ReXavatars: Evaluating Generated Patient Images for Clinical Realism and Reliability

Oishi Banerjee, Alexandra N. Willauer, Pranav Rajpurkar
Proceedings of the 11th Machine Learning for Healthcare Conference, PMLR 340:241-261, 2026.

Abstract

Generative image models enable the creation of realistic “patient avatars" that can undergo virtual examinations, with potential applications in medical education and AI agent benchmarking. We evaluate Gemini-3’s ability to generate such avatars using clinician-verified prompts and clinician assessment of photorealism, composition, and clinical accuracy. We find that images are highly realistic and well-composed, often achieving an impressive level of structural and textural detail. We further find that Gemini-3 can maintain anatomical and environmental consistency across viewpoints or poses in 60 and 70% of cases respectively, a key requirement for interactive scenarios such as virtual examinations. Despite these strengths, clinical accuracy is inconsistent. When testing without additional constraints, 45 of 80 pathologies tested are generated without errors; performance degrades when we increase task complexity by requiring specific presentations or combinations of pathologies. Additionally, we find notable weaknesses in numerical reasoning tasks, with a 45% success rate on tasks requiring accurate counting. These results highlight the promise of patient avatars as a new interface for education and evaluation, while showing directions to improve clinical accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v340-banerjee26a, title = {ReXavatars: Evaluating Generated Patient Images for Clinical Realism and Reliability}, author = {Banerjee, Oishi and Willauer, Alexandra N. and Rajpurkar, Pranav}, booktitle = {Proceedings of the 11th Machine Learning for Healthcare Conference}, pages = {241--261}, year = {2026}, editor = {Krishnan, Rahul G. and van Amsterdam, Wouter A. C. and Chopra, Sumit and Overgaard, Shauna and Hughes, Michael and Ötleş, Erkin and Shen, Yiqiu and Shanmugam, Divya and Nayan, Madhur and Engelhard, Matthew and Fackler, Jim and Oberst, Michael}, volume = {340}, series = {Proceedings of Machine Learning Research}, month = {12--14 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v340/main/assets/banerjee26a/banerjee26a.pdf}, url = {https://proceedings.mlr.press/v340/banerjee26a.html}, abstract = {Generative image models enable the creation of realistic “patient avatars" that can undergo virtual examinations, with potential applications in medical education and AI agent benchmarking. We evaluate Gemini-3’s ability to generate such avatars using clinician-verified prompts and clinician assessment of photorealism, composition, and clinical accuracy. We find that images are highly realistic and well-composed, often achieving an impressive level of structural and textural detail. We further find that Gemini-3 can maintain anatomical and environmental consistency across viewpoints or poses in 60 and 70% of cases respectively, a key requirement for interactive scenarios such as virtual examinations. Despite these strengths, clinical accuracy is inconsistent. When testing without additional constraints, 45 of 80 pathologies tested are generated without errors; performance degrades when we increase task complexity by requiring specific presentations or combinations of pathologies. Additionally, we find notable weaknesses in numerical reasoning tasks, with a 45% success rate on tasks requiring accurate counting. These results highlight the promise of patient avatars as a new interface for education and evaluation, while showing directions to improve clinical accuracy.} }
Endnote
%0 Conference Paper %T ReXavatars: Evaluating Generated Patient Images for Clinical Realism and Reliability %A Oishi Banerjee %A Alexandra N. Willauer %A Pranav Rajpurkar %B Proceedings of the 11th Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2026 %E Rahul G. Krishnan %E Wouter A. C. van Amsterdam %E Sumit Chopra %E Shauna Overgaard %E Michael Hughes %E Erkin Ötleş %E Yiqiu Shen %E Divya Shanmugam %E Madhur Nayan %E Matthew Engelhard %E Jim Fackler %E Michael Oberst %F pmlr-v340-banerjee26a %I PMLR %P 241--261 %U https://proceedings.mlr.press/v340/banerjee26a.html %V 340 %X Generative image models enable the creation of realistic “patient avatars" that can undergo virtual examinations, with potential applications in medical education and AI agent benchmarking. We evaluate Gemini-3’s ability to generate such avatars using clinician-verified prompts and clinician assessment of photorealism, composition, and clinical accuracy. We find that images are highly realistic and well-composed, often achieving an impressive level of structural and textural detail. We further find that Gemini-3 can maintain anatomical and environmental consistency across viewpoints or poses in 60 and 70% of cases respectively, a key requirement for interactive scenarios such as virtual examinations. Despite these strengths, clinical accuracy is inconsistent. When testing without additional constraints, 45 of 80 pathologies tested are generated without errors; performance degrades when we increase task complexity by requiring specific presentations or combinations of pathologies. Additionally, we find notable weaknesses in numerical reasoning tasks, with a 45% success rate on tasks requiring accurate counting. These results highlight the promise of patient avatars as a new interface for education and evaluation, while showing directions to improve clinical accuracy.
APA
Banerjee, O., Willauer, A.N. & Rajpurkar, P.. (2026). ReXavatars: Evaluating Generated Patient Images for Clinical Realism and Reliability. Proceedings of the 11th Machine Learning for Healthcare Conference, in Proceedings of Machine Learning Research 340:241-261 Available from https://proceedings.mlr.press/v340/banerjee26a.html.

Related Material