2023
Hazem Mohammad; Sameer Kishore; and Judhi Prasetyo
Developing an Object Detection and Gripping Mechanism Algorithm using Machine Learning
Journal Article
In: Journal of Applied Science and Advanced Engineering, vol. 1, iss. 2, pp. 47-54, 2023.
@article{Mohammad2023,
title = {Developing an Object Detection and Gripping Mechanism Algorithm using Machine Learning},
author = {Hazem Mohammad and Sameer Kishore and and Judhi Prasetyo},
doi = {https://doi.org/10.59097/jasae.v1i2.15},
year = {2023},
date = {2023-09-27},
urldate = {2023-09-27},
journal = {Journal of Applied Science and Advanced Engineering},
volume = {1},
issue = {2},
pages = {47-54},
abstract = {Localizing and recognition of objects are critical problems for indoor manipulation tasks. This paper describes an algorithm based on computer vision and machine learning that does object detection and gripping tasks. Detection of objects is carried out using a combination of a camera and depth sensor using a Kinect v1 depth sensor. Moreover, machine learning algorithms (YOLO) are used for computer vision. The project presents a method that allows the Kinect sensor to detect objects' 3D location. At the same time, it is attached to any robotic arm base, allowing for a more versatile and compact solution to be used in stationary places using industrial robot arms or mobile robots. The results show an error of locating an object to be 5 mm. and more than 70% confidence in detecting objects correctly. There are many possibilities in which this project can be used, such as in industrial fields, to sort, load, and unload different kinds of objects based on their type, size, and shape. In agriculture fields, to collect or sort different kinds of fruits, in kitchens and cafes where sorting objects like cups, bottles, and cans can occur. Also, this project can be added to mobile robots to do indoor human services or collect trash from different places.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Localizing and recognition of objects are critical problems for indoor manipulation tasks. This paper describes an algorithm based on computer vision and machine learning that does object detection and gripping tasks. Detection of objects is carried out using a combination of a camera and depth sensor using a Kinect v1 depth sensor. Moreover, machine learning algorithms (YOLO) are used for computer vision. The project presents a method that allows the Kinect sensor to detect objects' 3D location. At the same time, it is attached to any robotic arm base, allowing for a more versatile and compact solution to be used in stationary places using industrial robot arms or mobile robots. The results show an error of locating an object to be 5 mm. and more than 70% confidence in detecting objects correctly. There are many possibilities in which this project can be used, such as in industrial fields, to sort, load, and unload different kinds of objects based on their type, size, and shape. In agriculture fields, to collect or sort different kinds of fruits, in kitchens and cafes where sorting objects like cups, bottles, and cans can occur. Also, this project can be added to mobile robots to do indoor human services or collect trash from different places.
Olga Khokhlova†; Nishtha Lamba†; Sameer Kishore
Journal Article
In: Frontiers in Education, vol. 8, pp. 1158132, 2023, ISSN: 2504-284X, (†These authors have contributed equally to this work and share first authorship).
@article{khokhlova8evaluating,
title = {Evaluating student evaluations: evidence of gender bias against women in higher education based on perceived learning and instructor personality},
author = {Olga Khokhlova† and Nishtha Lamba† and Sameer Kishore},
editor = {Ana M. Vernia-Carrasco},
url = {https://www.frontiersin.org/articles/10.3389/feduc.2023.1158132},
doi = {10.3389/feduc.2023.1158132},
issn = {2504-284X},
year = {2023},
date = {2023-07-27},
urldate = {2023-07-27},
journal = {Frontiers in Education},
volume = {8},
pages = {1158132},
abstract = {Given student evaluations are an integral part of academic employment and progression in higher education, it is crucial to explore various biases amongst students that may influence their ratings. Several studies report a clear gender bias in student evaluation where male instructors receive significantly higher ratings as compared to female instructors. However, there is very limited research about gender biases in underrepresented samples such as South Asia and the Middle East. We examined whether perception of male and female instructors differed in terms of how they facilitate learning and level of engagement, using an experimental design. Six hundred and seventy-one university students were asked to watch a video of an online lecture on psychology, delivered by either a male or female lecturer, after which they were asked to evaluate their experience and instructor personality characteristics. To ensure consistency across content, tone, delivery, environment, and overall appearance, photorealistic 3D avatars were used to deliver the lectures. Only gender as a factor was manipulated. Given the racial representation in the region, a total of four videos were developed representing males (n = 317) and females (n = 354) of White and South Asian race. Overall, male instructors scored significantly higher in variables representing personality characteristics such as enthusiasm and expressiveness compared to female instructors. Participants did not however view male and female instructors to be different in terms of presentation and subject knowledge. Findings related to facilitating learning suggest that male instructors were perceived to have made instructions more interesting, kept participants' attention for longer, and were more interesting compared to female instructors. In terms of engagement, male instructors were perceived to be more expressive, enthusiastic, and entertaining, compared to female instructors. Given the experimental design, these findings can clearly be attributed to gender bias, which is also in line with previous research. With an underrepresented sample, an online platform delivery, and inclusion of multiple races, these findings significantly add value to the current literature regarding gender stereotypes in higher education. The results are even more concerning as they provide strong evidence of gender bias which may contribute to subconscious discrimination against women academics in the region.},
note = {†These authors have contributed equally to this work and share first authorship},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Given student evaluations are an integral part of academic employment and progression in higher education, it is crucial to explore various biases amongst students that may influence their ratings. Several studies report a clear gender bias in student evaluation where male instructors receive significantly higher ratings as compared to female instructors. However, there is very limited research about gender biases in underrepresented samples such as South Asia and the Middle East. We examined whether perception of male and female instructors differed in terms of how they facilitate learning and level of engagement, using an experimental design. Six hundred and seventy-one university students were asked to watch a video of an online lecture on psychology, delivered by either a male or female lecturer, after which they were asked to evaluate their experience and instructor personality characteristics. To ensure consistency across content, tone, delivery, environment, and overall appearance, photorealistic 3D avatars were used to deliver the lectures. Only gender as a factor was manipulated. Given the racial representation in the region, a total of four videos were developed representing males (n = 317) and females (n = 354) of White and South Asian race. Overall, male instructors scored significantly higher in variables representing personality characteristics such as enthusiasm and expressiveness compared to female instructors. Participants did not however view male and female instructors to be different in terms of presentation and subject knowledge. Findings related to facilitating learning suggest that male instructors were perceived to have made instructions more interesting, kept participants' attention for longer, and were more interesting compared to female instructors. In terms of engagement, male instructors were perceived to be more expressive, enthusiastic, and entertaining, compared to female instructors. Given the experimental design, these findings can clearly be attributed to gender bias, which is also in line with previous research. With an underrepresented sample, an online platform delivery, and inclusion of multiple races, these findings significantly add value to the current literature regarding gender stereotypes in higher education. The results are even more concerning as they provide strong evidence of gender bias which may contribute to subconscious discrimination against women academics in the region.
Nishtha Lamba; Sameer Kishore; Olga Khokhlova
Who Is More ‘Human-Like’? Exploring Gender and Race in Perception of Avatars in an Online Classroom
Conference
11th European Conference on Education, London, UK, 2023.
@conference{LambaHumanLike2023,
title = {Who Is More ‘Human-Like’? Exploring Gender and Race in Perception of Avatars in an Online Classroom},
author = {Nishtha Lamba and Sameer Kishore and Olga Khokhlova },
url = {https://www.youtube.com/watch?v=x_EAt1o3pfc},
year = {2023},
date = {2023-07-16},
urldate = {2023-07-16},
booktitle = {11th European Conference on Education},
address = {London, UK},
abstract = {The use of avatars in online and immersive environments plays a crucial role in shaping user experience. Given the recent shift towards remote interaction using online platforms, especially in the context of higher education, it is vital to study perceptions, attitudes, and biases based on avatar characteristics. The aim of this study is to explore perception of instructors represented by avatars in terms of realism and behaviour in an online undergraduate classroom. Using a between-group design, one of the four matched videos (Male x South Asian; Female x South Asian; Male x White; Female x White) showcasing a photorealistic avatar teaching online using a Zoom platform were shown to 327 UAE-based participants (Gender: M = 76, F = 246, Other = 5; Race: SA = 214, Other = 113). Five 2x2 ANOVA were administered on the items of subscale ‘Human-like’ from the Agent Persona questionnaire. Male avatars were perceived to have more personality and were more human-like compared to female avatars. South Asian avatars scored higher on items related to personality, natural emotions, being human-like, having natural movements, and showing emotions, compared to White avatars. Our results show that all avatars received consistently high ratings across dimensions related to human-like qualities. However, there was greater in-group preference based on race compared to gender. While these findings are encouraging, discrepancies in ratings indicate the existence of potential biases or preferences. Another notable contribution lies in the use of South Asian avatars, which have received little attention in
previous research.},
keywords = {},
pubstate = {published},
tppubtype = {conference}
}
The use of avatars in online and immersive environments plays a crucial role in shaping user experience. Given the recent shift towards remote interaction using online platforms, especially in the context of higher education, it is vital to study perceptions, attitudes, and biases based on avatar characteristics. The aim of this study is to explore perception of instructors represented by avatars in terms of realism and behaviour in an online undergraduate classroom. Using a between-group design, one of the four matched videos (Male x South Asian; Female x South Asian; Male x White; Female x White) showcasing a photorealistic avatar teaching online using a Zoom platform were shown to 327 UAE-based participants (Gender: M = 76, F = 246, Other = 5; Race: SA = 214, Other = 113). Five 2x2 ANOVA were administered on the items of subscale ‘Human-like’ from the Agent Persona questionnaire. Male avatars were perceived to have more personality and were more human-like compared to female avatars. South Asian avatars scored higher on items related to personality, natural emotions, being human-like, having natural movements, and showing emotions, compared to White avatars. Our results show that all avatars received consistently high ratings across dimensions related to human-like qualities. However, there was greater in-group preference based on race compared to gender. While these findings are encouraging, discrepancies in ratings indicate the existence of potential biases or preferences. Another notable contribution lies in the use of South Asian avatars, which have received little attention in
previous research.
previous research.
2022
Nishtha Lamba; Sameer Kishore; Olga Khokhlova
Conference
10th European Conference on Education, London, UK, 2022.
@conference{Lamba2022,
title = {Using Virtual Avatars To Explore the Impact of Instructors' Gender and Race on Perceived Learning in an Online Classroom},
author = {Nishtha Lamba and Sameer Kishore and Olga Khokhlova},
url = {https://www.youtube.com/watch?v=x_EAt1o3pfc},
year = {2022},
date = {2022-07-16},
urldate = {2022-07-16},
booktitle = {10th European Conference on Education},
address = {London, UK},
abstract = {There is substantial evidence that students’ perceptions and evaluations of instructors are influenced by their expectations and implicit biases, which further impacts their learning and motivation in higher education.
We have designed a study with an experimental design to examine the impact of A) gender (male, female) and race (South Asian, White) on perceived learning and interest in further learning from the same lecturer. 331 UAE-based students, respectively, were shown a video of a virtual instructor delivering a short online lecture. The avatars’ characteristics were manipulated one at a time while keeping content, tone, delivery, environment, and overall appearance consistent.
The results of the study show that gender or race did not significantly facilitate learning. Participants significantly preferred to study more from male instructors compared to female instructors F (1, 327) = 5.74, p = .01 and from South Asian instructors compared to White instructors F (1, 327) = 4.54, p = .03. Our results are in line with literature - Male instructors were rated higher than female instructors for several factors. Given there is limited information in the context of Middle East and South Asia, findings highlight differences specific to the region.},
keywords = {},
pubstate = {published},
tppubtype = {conference}
}
There is substantial evidence that students’ perceptions and evaluations of instructors are influenced by their expectations and implicit biases, which further impacts their learning and motivation in higher education.
We have designed a study with an experimental design to examine the impact of A) gender (male, female) and race (South Asian, White) on perceived learning and interest in further learning from the same lecturer. 331 UAE-based students, respectively, were shown a video of a virtual instructor delivering a short online lecture. The avatars’ characteristics were manipulated one at a time while keeping content, tone, delivery, environment, and overall appearance consistent.
The results of the study show that gender or race did not significantly facilitate learning. Participants significantly preferred to study more from male instructors compared to female instructors F (1, 327) = 5.74, p = .01 and from South Asian instructors compared to White instructors F (1, 327) = 4.54, p = .03. Our results are in line with literature - Male instructors were rated higher than female instructors for several factors. Given there is limited information in the context of Middle East and South Asia, findings highlight differences specific to the region.
We have designed a study with an experimental design to examine the impact of A) gender (male, female) and race (South Asian, White) on perceived learning and interest in further learning from the same lecturer. 331 UAE-based students, respectively, were shown a video of a virtual instructor delivering a short online lecture. The avatars’ characteristics were manipulated one at a time while keeping content, tone, delivery, environment, and overall appearance consistent.
The results of the study show that gender or race did not significantly facilitate learning. Participants significantly preferred to study more from male instructors compared to female instructors F (1, 327) = 5.74, p = .01 and from South Asian instructors compared to White instructors F (1, 327) = 4.54, p = .03. Our results are in line with literature - Male instructors were rated higher than female instructors for several factors. Given there is limited information in the context of Middle East and South Asia, findings highlight differences specific to the region.
2021
Sameer Kishore; Bernhard Spanlang; Guillermo Iruretagoyena; Shivashankar Halan; Dalila Szostak; Mel Slater
Journal Article
In: PRESENCE: Virtual and Augmented Reality, vol. 28, pp. 5-27, 2021, ISSN: 1054-7460.
@article{Kishore2021,
title = {A Virtual Reality Embodiment Technique to Enhance Helping Behavior of Police Towards a Victim of Police Racial Aggression},
author = {Sameer Kishore and Bernhard Spanlang and Guillermo Iruretagoyena and Shivashankar Halan and Dalila Szostak and Mel Slater},
url = {https://www.youtube.com/watch?v=tb9QAUkZWic},
doi = {10.1162/pres_a_00339},
issn = {1054-7460},
year = {2021},
date = {2021-10-29},
urldate = {2021-10-29},
journal = {PRESENCE: Virtual and Augmented Reality},
volume = {28},
pages = {5-27},
abstract = {There is an alarming level of violence by police in the US toward African Americans. Although this may be rooted in explicit racial bias, the more intractable problem is overcoming implicit bias, bias that is non-conscious but demonstrated in actual behavior. If bias is implicit, it is difficult to change through explicit methods that attempt to change attitudes. We carried out a study using virtual reality (VR) with 38 officers in a US police department, who took part in an interrogation of an African American suspect alongside an officer who was racially abusive toward the suspect. Seventeen of the participants witnessed the interview again from a third person perspective (Observer) and 21 from the embodied perspective of the suspect, now a victim of the interrogation (Victim condition), having been assigned randomly to these two groups. Some weeks later, all witnessed aggression by an officer toward an African American man in a virtual cafe scenario. The results show that the actions of those who had been in the Victim condition were coded as being more helpful toward the victim than those in the Observer condition. We argue that such VR exposures operate at the experiential and implicit level rather than the explicit, and hence are more likely to be effective in combating aggression rooted in implicit bias.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
There is an alarming level of violence by police in the US toward African Americans. Although this may be rooted in explicit racial bias, the more intractable problem is overcoming implicit bias, bias that is non-conscious but demonstrated in actual behavior. If bias is implicit, it is difficult to change through explicit methods that attempt to change attitudes. We carried out a study using virtual reality (VR) with 38 officers in a US police department, who took part in an interrogation of an African American suspect alongside an officer who was racially abusive toward the suspect. Seventeen of the participants witnessed the interview again from a third person perspective (Observer) and 21 from the embodied perspective of the suspect, now a victim of the interrogation (Victim condition), having been assigned randomly to these two groups. Some weeks later, all witnessed aggression by an officer toward an African American man in a virtual cafe scenario. The results show that the actions of those who had been in the Victim condition were coded as being more helpful toward the victim than those in the Observer condition. We argue that such VR exposures operate at the experiential and implicit level rather than the explicit, and hence are more likely to be effective in combating aggression rooted in implicit bias.
2019
Laura Aymerich-Franch; Sameer Kishore; Mel Slater
Journal Article
In: International Journal of Social Robotics, vol. 12, iss. 1, pp. 217-226, 2019, ISSN: 1875-4805.
@article{Aymerich-Franch2019,
title = {When Your Robot Avatar Misbehaves You Are Likely to Apologize: An Exploration of Guilt During Robot Embodiment},
author = {Laura Aymerich-Franch and Sameer Kishore and Mel Slater},
url = {https://link.springer.com/article/10.1007/s12369-019-00556-5
https://www.youtube.com/watch?v=FTJieCy1kos},
doi = {10.1007/s12369-019-00556-5},
issn = {1875-4805},
year = {2019},
date = {2019-05-09},
urldate = {2019-05-09},
journal = {International Journal of Social Robotics},
volume = {12},
issue = {1},
pages = {217-226},
abstract = {Would people feel guilty if their robot avatar acted autonomously to harm someone? We examined the experience of guilt during robot avatar embodiment, a form of embodiment where the participants experience the body of a humanoid robot as if it were their own. In particular, we analyzed what happens when a robot avatar spontaneously verbally abuses someone during a conversation using the participant’s voice, without this being the intention of the participant. In a 2 × 2 between-subjects experimental design, participants embodied a humanoid robot that added either offensive or neutral words during a conversation with a confederate, and had control over the robot’s movements or not (synch. vs. asynch.). We found that guilt and shame were positively associated with offensive words and that apologizing and verbal repair were positively related to guilt. Also, body ownership was moderately associated to apologizing and verbal repair. The results suggest that people may feel guilty for the actions of their robot avatars even if they are not the real agents of these actions. The work highlights the importance of examining the moral and legal aspects related to robot embodiment technologies.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Would people feel guilty if their robot avatar acted autonomously to harm someone? We examined the experience of guilt during robot avatar embodiment, a form of embodiment where the participants experience the body of a humanoid robot as if it were their own. In particular, we analyzed what happens when a robot avatar spontaneously verbally abuses someone during a conversation using the participant’s voice, without this being the intention of the participant. In a 2 × 2 between-subjects experimental design, participants embodied a humanoid robot that added either offensive or neutral words during a conversation with a confederate, and had control over the robot’s movements or not (synch. vs. asynch.). We found that guilt and shame were positively associated with offensive words and that apologizing and verbal repair were positively related to guilt. Also, body ownership was moderately associated to apologizing and verbal repair. The results suggest that people may feel guilty for the actions of their robot avatars even if they are not the real agents of these actions. The work highlights the importance of examining the moral and legal aspects related to robot embodiment technologies.
2018
Daniel Freeman; Polly Haselton; Jason Freeman; Bernhard Spanlang; Sameer Kishore; Emily Albery; Megan Denne; Poppy Brown; Mel Slater; Alicia Nickless
Journal Article
In: The Lancet Psychiatry, vol. 5, iss. 8, pp. 625-632, 2018, ISSN: 2215-0374.
@article{Freeman2018,
title = {Automated psychological therapy using immersive virtual reality for treatment of fear of heights: a single-blind, parallel-group, randomised controlled trial},
author = {Daniel Freeman and Polly Haselton and Jason Freeman and Bernhard Spanlang and Sameer Kishore and Emily Albery and Megan Denne and Poppy Brown and Mel Slater and Alicia Nickless},
url = {https://www.independent.co.uk/news/health/virtual-reality-phobia-fear-heights-mental-health-treatment-oxford-university-lancet-a8442696.html},
doi = {10.1016/S2215-0366(18)30226-8},
issn = {2215-0374},
year = {2018},
date = {2018-08-01},
urldate = {2018-08-01},
journal = {The Lancet Psychiatry},
volume = {5},
issue = {8},
pages = {625-632},
abstract = {Background
Engaging, interactive, and automated virtual reality (VR) treatments might help solve the unmet needs of individuals with mental health disorders. We tested the efficacy of an automated cognitive intervention for fear of heights guided by an avatar virtual coach (animated using motion and voice capture of an actor) in VR and delivered with the latest consumer equipment.
Methods
We did a randomised trial of automated VR versus usual care. We recruited adults aged older than 18 years with a fear of heights by radio advertisements in Oxfordshire, UK. We diagnosed fear of heights if participants scored more than 29 on the Heights Interpretation Questionnaire (HIQ). We randomly allocated participants by computer in a 1:1 ratio to either automated VR delivered in roughly six 30-min sessions administered about two to three times a week over a 2-week period (intervention group) or to usual care (control group). Randomisation was stratified by severity of fear of heights. The research team, who were unaware of the random allocation, administered three fear-of-height assessments, at baseline (0 weeks), at the end of treatment (2 weeks), and at follow-up (4 weeks). The primary outcome measure was HIQ score (range 16–80, with higher scores indicating greater severity). This trial is registered with the ISRCTN registry, number ISRCTN11898283.
Findings
Between Nov 25, 2017, and Feb 27, 2018, 100 individuals were enrolled and underwent randomisation, of whom 49 were assigned to the VR treatment group and 51 to the control group. All participants completed the 4-week follow-up. The mean total treatment time in VR was 124·43 min (SD 34·23). Compared with participants in the control group, the VR treatment reduced fear of heights at the end of treatment (mean change score −24·5 [SD 13·1] in the VR group vs −1·2 [7·3] in the control group; adjusted difference −24·0, 95% CI −27·7 to −20·3; Cohen's d=2·0; p<0·0001). The benefit was maintained at follow-up (mean change score −25·1 [SD 13·9] in the VR group vs −1·5 [7·8] in the control group; adjusted difference −24·3, 95% CI −27·9 to −20·6; Cohen's d=2·0; p<0·0001). The number needed to treat to at least halve the fear of heights was 1·3. No adverse events were reported.
Interpretation
Psychological therapy delivered automatically by a VR coach can produce large clinical benefits. Evidence-based VR treatments have the potential to greatly increase treatment provision for mental health disorders.
Funding
Oxford VR, and the National Institute of Health Research Oxford Health Biomedical Research Centre.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Background
Engaging, interactive, and automated virtual reality (VR) treatments might help solve the unmet needs of individuals with mental health disorders. We tested the efficacy of an automated cognitive intervention for fear of heights guided by an avatar virtual coach (animated using motion and voice capture of an actor) in VR and delivered with the latest consumer equipment.
Methods
We did a randomised trial of automated VR versus usual care. We recruited adults aged older than 18 years with a fear of heights by radio advertisements in Oxfordshire, UK. We diagnosed fear of heights if participants scored more than 29 on the Heights Interpretation Questionnaire (HIQ). We randomly allocated participants by computer in a 1:1 ratio to either automated VR delivered in roughly six 30-min sessions administered about two to three times a week over a 2-week period (intervention group) or to usual care (control group). Randomisation was stratified by severity of fear of heights. The research team, who were unaware of the random allocation, administered three fear-of-height assessments, at baseline (0 weeks), at the end of treatment (2 weeks), and at follow-up (4 weeks). The primary outcome measure was HIQ score (range 16–80, with higher scores indicating greater severity). This trial is registered with the ISRCTN registry, number ISRCTN11898283.
Findings
Between Nov 25, 2017, and Feb 27, 2018, 100 individuals were enrolled and underwent randomisation, of whom 49 were assigned to the VR treatment group and 51 to the control group. All participants completed the 4-week follow-up. The mean total treatment time in VR was 124·43 min (SD 34·23). Compared with participants in the control group, the VR treatment reduced fear of heights at the end of treatment (mean change score −24·5 [SD 13·1] in the VR group vs −1·2 [7·3] in the control group; adjusted difference −24·0, 95% CI −27·7 to −20·3; Cohen's d=2·0; p<0·0001). The benefit was maintained at follow-up (mean change score −25·1 [SD 13·9] in the VR group vs −1·5 [7·8] in the control group; adjusted difference −24·3, 95% CI −27·9 to −20·6; Cohen's d=2·0; p<0·0001). The number needed to treat to at least halve the fear of heights was 1·3. No adverse events were reported.
Interpretation
Psychological therapy delivered automatically by a VR coach can produce large clinical benefits. Evidence-based VR treatments have the potential to greatly increase treatment provision for mental health disorders.
Funding
Oxford VR, and the National Institute of Health Research Oxford Health Biomedical Research Centre.
Engaging, interactive, and automated virtual reality (VR) treatments might help solve the unmet needs of individuals with mental health disorders. We tested the efficacy of an automated cognitive intervention for fear of heights guided by an avatar virtual coach (animated using motion and voice capture of an actor) in VR and delivered with the latest consumer equipment.
Methods
We did a randomised trial of automated VR versus usual care. We recruited adults aged older than 18 years with a fear of heights by radio advertisements in Oxfordshire, UK. We diagnosed fear of heights if participants scored more than 29 on the Heights Interpretation Questionnaire (HIQ). We randomly allocated participants by computer in a 1:1 ratio to either automated VR delivered in roughly six 30-min sessions administered about two to three times a week over a 2-week period (intervention group) or to usual care (control group). Randomisation was stratified by severity of fear of heights. The research team, who were unaware of the random allocation, administered three fear-of-height assessments, at baseline (0 weeks), at the end of treatment (2 weeks), and at follow-up (4 weeks). The primary outcome measure was HIQ score (range 16–80, with higher scores indicating greater severity). This trial is registered with the ISRCTN registry, number ISRCTN11898283.
Findings
Between Nov 25, 2017, and Feb 27, 2018, 100 individuals were enrolled and underwent randomisation, of whom 49 were assigned to the VR treatment group and 51 to the control group. All participants completed the 4-week follow-up. The mean total treatment time in VR was 124·43 min (SD 34·23). Compared with participants in the control group, the VR treatment reduced fear of heights at the end of treatment (mean change score −24·5 [SD 13·1] in the VR group vs −1·2 [7·3] in the control group; adjusted difference −24·0, 95% CI −27·7 to −20·3; Cohen's d=2·0; p<0·0001). The benefit was maintained at follow-up (mean change score −25·1 [SD 13·9] in the VR group vs −1·5 [7·8] in the control group; adjusted difference −24·3, 95% CI −27·9 to −20·6; Cohen's d=2·0; p<0·0001). The number needed to treat to at least halve the fear of heights was 1·3. No adverse events were reported.
Interpretation
Psychological therapy delivered automatically by a VR coach can produce large clinical benefits. Evidence-based VR treatments have the potential to greatly increase treatment provision for mental health disorders.
Funding
Oxford VR, and the National Institute of Health Research Oxford Health Biomedical Research Centre.
Domna Banakou; Sameer Kishore; Mel Slater
Journal Article
In: Frontiers in psychology, vol. 9, 2018, ISSN: 1664-1078.
@article{Banakou2018,
title = {Virtually being Einstein results in an improvement in cognitive task performance and a decrease in age bias},
author = {Domna Banakou and Sameer Kishore and Mel Slater},
url = {https://www.frontiersin.org/articles/10.3389/fpsyg.2018.00917/full
https://www.youtube.com/watch?v=TAEM5OlFbnw},
doi = {10.3389/fpsyg.2018.00917},
issn = {1664-1078},
year = {2018},
date = {2018-06-11},
urldate = {2018-06-11},
journal = {Frontiers in psychology},
volume = {9},
abstract = {The brain's body representation is amenable to rapid change, even though we tend to think of our bodies as relatively fixed and stable. For example, it has been shown that a life-sized body perceived in virtual reality as substituting the participant's real body, can be felt as if it were their own, and that the body type can induce perceptual, attitudinal and behavioral changes. Here we show that changes can also occur in cognitive processing and specifically, executive functioning. Fifteen male participants were embodied in a virtual body that signifies super-intelligence (Einstein) and 15 in a (Normal) virtual body of similar age to their own. The Einstein body participants performed better on a cognitive task than the Normal body, considering prior cognitive ability (IQ), with the improvement greatest for those with low self-esteem. Einstein embodiment also reduced implicit bias against older people. Hence virtual body ownership may additionally be used to enhance executive functioning.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The brain's body representation is amenable to rapid change, even though we tend to think of our bodies as relatively fixed and stable. For example, it has been shown that a life-sized body perceived in virtual reality as substituting the participant's real body, can be felt as if it were their own, and that the body type can induce perceptual, attitudinal and behavioral changes. Here we show that changes can also occur in cognitive processing and specifically, executive functioning. Fifteen male participants were embodied in a virtual body that signifies super-intelligence (Einstein) and 15 in a (Normal) virtual body of similar age to their own. The Einstein body participants performed better on a cognitive task than the Normal body, considering prior cognitive ability (IQ), with the improvement greatest for those with low self-esteem. Einstein embodiment also reduced implicit bias against older people. Hence virtual body ownership may additionally be used to enhance executive functioning.
Sameer Kishore; Xavi Navarro; Eva Dominguez; Nonny de la Peña; Mel Slater
Beaming into the News: A System for and Case Study of Tele-Immersive Journalism
Journal Article
In: IEEE Computer Graphics and Applications, vol. 38, iss. 2, pp. 89-101, 2018, ISSN: 0272-1716.
@article{Kishore2016,
title = {Beaming into the News: A System for and Case Study of Tele-Immersive Journalism},
author = {Sameer Kishore and Xavi Navarro and Eva Dominguez and Nonny de la Peña and Mel Slater},
url = {https://www.youtube.com/watch?v=ry6dmWB34qI},
doi = {10.1109/MCG.2016.44},
issn = {0272-1716},
year = {2018},
date = {2018-04-12},
urldate = {2018-04-12},
journal = {IEEE Computer Graphics and Applications},
volume = {38},
issue = {2},
pages = {89-101},
abstract = {We show how a combination of virtual reality and robotics can be used to beam a physical representation of a person to a distant location, and describe an application of this system in the context of journalism. Full body motion capture data of a person is streamed and mapped in real time, onto the limbs of a humanoid robot present at the remote location. A pair of cameras in the robot's 'eyes' stream stereoscopic video back to the HMD worn by the visitor, and a two-way audio connection allows the visitor to talk to people in the remote destination. By fusing the multisensory data of the visitor with the robot, the visitor's 'consciousness' is transformed to the robot's body. This system was used by a journalist to interview a neuroscientist and a chef 900 miles distant, about food for the brain, resulting in an article published in the popular press.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We show how a combination of virtual reality and robotics can be used to beam a physical representation of a person to a distant location, and describe an application of this system in the context of journalism. Full body motion capture data of a person is streamed and mapped in real time, onto the limbs of a humanoid robot present at the remote location. A pair of cameras in the robot's 'eyes' stream stereoscopic video back to the HMD worn by the visitor, and a two-way audio connection allows the visitor to talk to people in the remote destination. By fusing the multisensory data of the visitor with the robot, the visitor's 'consciousness' is transformed to the robot's body. This system was used by a journalist to interview a neuroscientist and a chef 900 miles distant, about food for the brain, resulting in an article published in the popular press.
2016
Sameer Kishore; Xavi Navarro; Pierre Bourdin; Keren Or-Berkers; Doron Friedman; Mel Slater
Journal Article
In: Frontiers in Robotics and AI, vol. 3, pp. 65, 2016, ISSN: 2296-9144.
@article{Kishore2016a,
title = {Multi-Destination Beaming: Apparently Being in Three Places at Once through Robotic and Virtual Embodiment},
author = {Sameer Kishore and Xavi Navarro and Pierre Bourdin and Keren Or-Berkers and Doron Friedman and Mel Slater},
url = {http://journal.frontiersin.org/article/10.3389/frobt.2016.00065/full
https://www.youtube.com/watch?v=oh1B6C3JggQ},
doi = {10.3389/frobt.2016.00065},
issn = {2296-9144},
year = {2016},
date = {2016-11-01},
urldate = {2016-11-01},
journal = {Frontiers in Robotics and AI},
volume = {3},
pages = {65},
abstract = {It has been shown that an illusion of ownership over an artificial limb or even an entire body can be induced in people through multisensory stimulation, providing evidence that the surrogate body is the person’s actual body. Such body ownership illusions (BOIs) have been shown to occur with virtual bodies, mannequins, and humanoid robots. In this study, we show the possibility of eliciting a full-BOI over not one, but multiple artificial bodies concurrently. We demonstrate this by describing a system that allowed a participant to inhabit and fully control two different humanoid robots located in two distinct places and a virtual body in immersive virtual reality, using real-time full-body tracking and two-way audio communication, thereby giving them the illusion of ownership over each of them. We implemented this by allowing the participant be embodied in any one surrogate body at a given moment and letting them instantaneously switch between them. While the participant was embodied in one of the bodies, a proxy system would track the locations currently unoccupied and would control their remote representation in order to continue performing the tasks in those locations in a logical fashion. To test the efficacy of this system, an exploratory study was carried out with a fully functioning setup with three destinations and a simplified version of the proxy for use in a social interaction. The results indicate that the system was physically and psychologically comfortable and was rated highly by participants in terms of usability. Additionally, feelings of BOI and agency were reported, which were not influenced by the type of body representation. The results provide us with clues regarding BOI with humanoid robots of different dimensions, along with insight about self-localization and multilocation.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
It has been shown that an illusion of ownership over an artificial limb or even an entire body can be induced in people through multisensory stimulation, providing evidence that the surrogate body is the person’s actual body. Such body ownership illusions (BOIs) have been shown to occur with virtual bodies, mannequins, and humanoid robots. In this study, we show the possibility of eliciting a full-BOI over not one, but multiple artificial bodies concurrently. We demonstrate this by describing a system that allowed a participant to inhabit and fully control two different humanoid robots located in two distinct places and a virtual body in immersive virtual reality, using real-time full-body tracking and two-way audio communication, thereby giving them the illusion of ownership over each of them. We implemented this by allowing the participant be embodied in any one surrogate body at a given moment and letting them instantaneously switch between them. While the participant was embodied in one of the bodies, a proxy system would track the locations currently unoccupied and would control their remote representation in order to continue performing the tasks in those locations in a logical fashion. To test the efficacy of this system, an exploratory study was carried out with a fully functioning setup with three destinations and a simplified version of the proxy for use in a social interaction. The results indicate that the system was physically and psychologically comfortable and was rated highly by participants in terms of usability. Additionally, feelings of BOI and agency were reported, which were not influenced by the type of body representation. The results provide us with clues regarding BOI with humanoid robots of different dimensions, along with insight about self-localization and multilocation.
Sameer Kishore
PhD Thesis
2016.
@phdthesis{Kishore2016Thesis,
title = {Robotic Embodiment: Developing a System for and Applications with Full Body Ownership of a Humanoid Robot},
author = {Sameer Kishore},
editor = {Universitat de Barcelona},
url = {http://hdl.handle.net/2445/100809},
year = {2016},
date = {2016-07-15},
urldate = {2016-07-15},
abstract = {It has been shown that with appropriate multisensory stimulation an illusion of owning an artificial object as part of their own body can be induced in people. Such body ownership illusions have been shown to occur with artificial limbs, such as rubber hands, and even entire artificial or virtual bodies. Although extensive research has been carried out regarding full body ownership illusions with mannequins and virtual bodies, few studies exist that apply this concept to humanoid robots. On the other hand, extensive research has been carried out with robots in terms of telepresence and remote manipulation of the robot, known as teleoperation. Combining these concepts would give rise to a highly immersive, embodied experience in a humanoid robot located at a remote physical location, which holds great potential in terms of real-world applications. In this thesis, we aim to apply this phenomenon of full body ownership illusions in the context of humanoid robots, and to develop real-world applications where this technology could be beneficial. More specifically, by relying on knowledge gained from previous studies regarding body ownership illusions, we investigated whether it is possible to elicit this illusion with a humanoid robot. In addition, we developed a system in the context of telepresence robots, where the participant is embodied in a humanoid robot that is present in a different physical location, and can use this robotic body to interact with the remote environment. To test the functionality of the system and to gain an understanding of body ownership illusions with robots, we carried out two experimental studies and one case-study of a demonstration of the system as a real-world application. In the Brain-Computer Interface versus Eye Tracker study, we used our system to investigate whether it was possible to induce a full body ownership illusion over a humanoid robot with a highly ‘robotic’ appearance. In addition, we compared two different abstract methods of control, a Steady-State Visually Evoked Potential (SSVEP) based Brain-Computer Interface and eye-tracking, in an immersive environment to drive the robot. This was done mainly as a motivation for developing a prototype of a system that could be used by disabled patients. Our results showed that a feeling of body ownership illusion and agency can be induced, even though the postures between participants and the embodied robot were incongruent (the participant was sitting, while the robot was standing). Additionally, both BCI and eye tracking were reported to be suitable methods of control, although the degree of body ownership illusion was influenced by the control method, with higher scores of ownership reported for the BCI condition. In the Tele-Immersive Journalism case study, we used the same system as above, but with the added capability of letting the participant control the robot body by moving their own body. Since in this case we provided synchronous visuomotor correlations with the robotic body we expected this to result in an even higher level of body ownership illusion. By making the robot body the source of their associated sensations we simulate a type of virtual teleportation. We applied this system successfully to the context of journalism, where a journalist could be embodied in a humanoid robot located in a remote destination and carry out interviews through their robotic body. We provide a case-study where the system was used by several journalists to report news about the system itself as well as for reporting other stories. In the Multi-Destination Beaming study, we extended the functionality of the system to include three destinations. The aim of the study was to investigate whether participants could cope with being in three places at same time, and embodied in three different surrogate bodies. We had two physical destinations with one robot in each, and a third virtual destination where the participant would be embodied in a virtual body. The results indicate that the system was physically and psychologically comfortable, and was rated highly by participants in terms of usability in real world. Additionally, high feelings of body ownership illusion and agency were reported, which were not influenced by the robot type. This provides us with clues regarding body ownership illusion with humanoid robots of different dimensions, along with insight about self-localisation and multilocation. Overall, our results show that it is possible to elicit a full body ownership illusion over humanoid robotic bodies. The studies presented here advance the current theoretical framework of body representation, agency and self-perception by providing information about various factors that may affect the illusion of body ownership, such as a highly robotic appearance of the artificial body, having indirect methods of control, or even being simultaneously embodied in three different bodies. Additionally, the setup described can also be used to great effect for highly immersive remote robotic embodiment applications, such as one demonstrated here in the field of journalism.},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
It has been shown that with appropriate multisensory stimulation an illusion of owning an artificial object as part of their own body can be induced in people. Such body ownership illusions have been shown to occur with artificial limbs, such as rubber hands, and even entire artificial or virtual bodies. Although extensive research has been carried out regarding full body ownership illusions with mannequins and virtual bodies, few studies exist that apply this concept to humanoid robots. On the other hand, extensive research has been carried out with robots in terms of telepresence and remote manipulation of the robot, known as teleoperation. Combining these concepts would give rise to a highly immersive, embodied experience in a humanoid robot located at a remote physical location, which holds great potential in terms of real-world applications. In this thesis, we aim to apply this phenomenon of full body ownership illusions in the context of humanoid robots, and to develop real-world applications where this technology could be beneficial. More specifically, by relying on knowledge gained from previous studies regarding body ownership illusions, we investigated whether it is possible to elicit this illusion with a humanoid robot. In addition, we developed a system in the context of telepresence robots, where the participant is embodied in a humanoid robot that is present in a different physical location, and can use this robotic body to interact with the remote environment. To test the functionality of the system and to gain an understanding of body ownership illusions with robots, we carried out two experimental studies and one case-study of a demonstration of the system as a real-world application. In the Brain-Computer Interface versus Eye Tracker study, we used our system to investigate whether it was possible to induce a full body ownership illusion over a humanoid robot with a highly ‘robotic’ appearance. In addition, we compared two different abstract methods of control, a Steady-State Visually Evoked Potential (SSVEP) based Brain-Computer Interface and eye-tracking, in an immersive environment to drive the robot. This was done mainly as a motivation for developing a prototype of a system that could be used by disabled patients. Our results showed that a feeling of body ownership illusion and agency can be induced, even though the postures between participants and the embodied robot were incongruent (the participant was sitting, while the robot was standing). Additionally, both BCI and eye tracking were reported to be suitable methods of control, although the degree of body ownership illusion was influenced by the control method, with higher scores of ownership reported for the BCI condition. In the Tele-Immersive Journalism case study, we used the same system as above, but with the added capability of letting the participant control the robot body by moving their own body. Since in this case we provided synchronous visuomotor correlations with the robotic body we expected this to result in an even higher level of body ownership illusion. By making the robot body the source of their associated sensations we simulate a type of virtual teleportation. We applied this system successfully to the context of journalism, where a journalist could be embodied in a humanoid robot located in a remote destination and carry out interviews through their robotic body. We provide a case-study where the system was used by several journalists to report news about the system itself as well as for reporting other stories. In the Multi-Destination Beaming study, we extended the functionality of the system to include three destinations. The aim of the study was to investigate whether participants could cope with being in three places at same time, and embodied in three different surrogate bodies. We had two physical destinations with one robot in each, and a third virtual destination where the participant would be embodied in a virtual body. The results indicate that the system was physically and psychologically comfortable, and was rated highly by participants in terms of usability in real world. Additionally, high feelings of body ownership illusion and agency were reported, which were not influenced by the robot type. This provides us with clues regarding body ownership illusion with humanoid robots of different dimensions, along with insight about self-localisation and multilocation. Overall, our results show that it is possible to elicit a full body ownership illusion over humanoid robotic bodies. The studies presented here advance the current theoretical framework of body representation, agency and self-perception by providing information about various factors that may affect the illusion of body ownership, such as a highly robotic appearance of the artificial body, having indirect methods of control, or even being simultaneously embodied in three different bodies. Additionally, the setup described can also be used to great effect for highly immersive remote robotic embodiment applications, such as one demonstrated here in the field of journalism.
2014
Sameer Kishore; Mar González-Franco; Christoph Hintemüller; Christoph Kapeller; Christoph Guger; Mel Slater; Kristopher J. Blom
Comparison of SSVEP BCI and Eye Tracking for Controlling a Humanoid Robot in a Social Environment
Journal Article
In: Presence: Teleoperators and Virtual Environments, vol. 23, iss. 3, pp. 242-252, 2014, ISBN: 978-0-1996-4446-9.
@article{Kishore2014,
title = {Comparison of SSVEP BCI and Eye Tracking for Controlling a Humanoid Robot in a Social Environment},
author = {Sameer Kishore and Mar González-Franco and Christoph Hintemüller and Christoph Kapeller and Christoph Guger and Mel Slater and Kristopher J. Blom},
url = {https://www.youtube.com/watch?v=iGurLgspQxA},
doi = {10.1162/PRES_a_00192},
isbn = {978-0-1996-4446-9},
year = {2014},
date = {2014-10-01},
urldate = {2014-10-01},
journal = {Presence: Teleoperators and Virtual Environments},
volume = {23},
issue = {3},
pages = {242-252},
abstract = {Recent advances in humanoid robot technologies have made it possible to inhabit a humanlike form located at a remote place. This allows the participant to interact with others in that space and experience the illusion that the participant is actually present in the remote space. Moreover, with these humanlike forms, it may be possible to induce a full-body ownership illusion, where the robot body is perceived to be one’s own. We show that it is possible to induce the full-body ownership illusion over a remote robotic body with a highly robotic appearance. Additionally, our results indi- cate that even with nonmanual control of a remote robotic body, it is possible to induce feelings of agency and illusions of body ownership. Two established control methods, an SSVEP-based BCI and eye tracking, were tested as a means of controlling the robot’s gesturing. Our experience and the results indicate that both methods are tractable for immersive control of a humanoid robot in a social telepresence setting.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Recent advances in humanoid robot technologies have made it possible to inhabit a humanlike form located at a remote place. This allows the participant to interact with others in that space and experience the illusion that the participant is actually present in the remote space. Moreover, with these humanlike forms, it may be possible to induce a full-body ownership illusion, where the robot body is perceived to be one’s own. We show that it is possible to induce the full-body ownership illusion over a remote robotic body with a highly robotic appearance. Additionally, our results indi- cate that even with nonmanual control of a remote robotic body, it is possible to induce feelings of agency and illusions of body ownership. Two established control methods, an SSVEP-based BCI and eye tracking, were tested as a means of controlling the robot’s gesturing. Our experience and the results indicate that both methods are tractable for immersive control of a humanoid robot in a social telepresence setting.
2013
Bernhard Spanlang; Xavi Navarro; Jean-Marie Normand; Sameer Kishore; Rodrigo Pizarro; Mel Slater
Real Time Whole Body Motion Mapping for Avatars and Robots
Conference
19th ACM Symposium on Virtual Reality Software and Technology, ACM Press, New York, New York, USA, 2013, ISBN: 978-1-4503-2379-6.
@conference{Spanlang2013,
title = {Real Time Whole Body Motion Mapping for Avatars and Robots},
author = {Bernhard Spanlang and Xavi Navarro and Jean-Marie Normand and Sameer Kishore and Rodrigo Pizarro and Mel Slater},
url = {http://dl.acm.org/citation.cfm?doid=2503713.2503747
https://www.youtube.com/watch?v=jJ3NdybR3dE},
doi = {10.1145/2503713.2503747},
isbn = {978-1-4503-2379-6},
year = {2013},
date = {2013-10-06},
urldate = {2013-10-06},
booktitle = {19th ACM Symposium on Virtual Reality Software and Technology},
pages = {175-178},
publisher = {ACM Press},
address = {New York, New York, USA},
abstract = {We describe a system that allows for controlling different robots and avatars from a real time motion stream. The underlying problem is that motion data from tracking systems is usually represented differently to the motion data required to drive an avatar or a robot: there may be different joints, motion may be represented by absolute joint positions and rotations or by a root position, bone lengths and relative rotations in the skeletal hierarchy. Our system resolves these issues by remapping in real time the tracked motion so that the avatar or robot performs motions that are visually close to those of the tracked person. The mapping can also be reconfigured interactively at run-time. We demonstrate the effectiveness of our system by case studies in which a tracked person is embodied as an avatar in immersive virtual reality or as a robot in a remote location. We show this with a variety of tracking systems, humanoid avatars and robots.},
keywords = {},
pubstate = {published},
tppubtype = {conference}
}
We describe a system that allows for controlling different robots and avatars from a real time motion stream. The underlying problem is that motion data from tracking systems is usually represented differently to the motion data required to drive an avatar or a robot: there may be different joints, motion may be represented by absolute joint positions and rotations or by a root position, bone lengths and relative rotations in the skeletal hierarchy. Our system resolves these issues by remapping in real time the tracked motion so that the avatar or robot performs motions that are visually close to those of the tracked person. The mapping can also be reconfigured interactively at run-time. We demonstrate the effectiveness of our system by case studies in which a tracked person is embodied as an avatar in immersive virtual reality or as a robot in a remote location. We show this with a variety of tracking systems, humanoid avatars and robots.