Avinab Saha

I am Research Scientist at Google Research, developing reward models and evaluation methods for visual quality in generative multimodal image and video foundation models.

I recently graduated with a Ph.D. from the Laboratory of Image and Video Engineering, at The University of Texas at Austin, where I was fortunate to be advised by Prof. Al Bovik. During my Ph.D., I focused on Perceptual Image and Video Quality Assessment, with a particular emphasis on emerging technologies such as cloud gaming and virtual reality applications.

Previously, I was a Research Engineer at Samsung Research where I worked on the development of cutting-edge, AI-driven, low-complexity real-time algorithms for image and video enhancement applications that were commercialized in Samsung Galaxy Mobile Phones and 8K Televisions. Prior to that in 2019, I received my Bachelor's degree in Electronics and Electrical Communications Engineering from IIT Kharagpur, where I worked at the Visual Information Processing Lab with Prof. Jayanta Mukhopadhyay on accelerating Deep Learning models.

Email / Google Scholar / GitHub / Twitter(X) / LinkedIn / Instagram

News

05/2025: Transitioned to a full-time Research Scientist role at Google Research.
04/2025: Ph.D. dissertation successfully defended. Officially Dr. Saha!
04/2025: Paper accepted in SIGGRAPH 2025. See you in Vancouver!
03/2025: Paper accepted in CVPR 2025.
11/2024: Passed PhD Progress Review!
08/2024: Journal Paper accepted in IEEE Transactions on Image Processing.
05/2024: Joined Google Research as Research Intern in Athena Team!
04/2024: Paper accepted in 3rd Explainable AI for Computer Vision (XAI4CV) Workshop, CVPR 2024.

Research & Development Experience

05/2025 - Present : Research Scientist, Google Research, Mountain View.
05/2024 - 04/2025 : Research Intern/Student Researcher, Google Research, Mountain View.
05/2022 - 08/2022 : PhD Research Intern, Apple, Cupertino in Display & Color Technologies Team.
06/2019 - 01/2021 : Research Engineer, Samsung Research in AI Visual Processing Lab.
05/2018 - 08/2018 : Research Intern, Samsung Research in AI Visual Processing Lab.

Research Ongoing/Under Review

HoloQA: Full Reference Video Quality Assessor of Rendered Human Avatars in Virtual Reality
Subjective and Objective Quality Assessment of 3D Rendered Human Avatars.
Under revision (accepted with minor revisions) at IEEE Transactions on Image Processing,

Peer Reviewed Research Articles

	FaceExpressions-70k: A Dataset of Perceived Expression Differences Avinab Saha, Yu-Chih Chen, Christian Häne, Jean-Charles Bazin, Ioannis Katsavounidis, Alexandre Chapiro, Alan C. Bovik ACM Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques (SIGGRAPH) 2025 html / paper / project page / database / video
	Focus-N-Fix: Region-Aware Fine-Tuning for Text-to-Image Generation Xiaoying Xing, Avinab Saha, Junfeng He, Susan Hao, Paul Vicol, Moonkyung Ryu, Gang Li, Sahil Singla, Sarah Young, Yinxiao Li, Feng Yang, Deepak Ramachandran IEEE/CVF Computer Vision and Pattern Recognition (CVPR)* 2025 arxiv / paper / video
	Subjective and Objective Quality Assessment of Rendered Human Avatar Videos in Virtual Reality Yu-Chih Chen, Avinab Saha, Alexandre Chapiro, Christian Häne, Jean-Charles Bazin, Bo Qiu, Stefano Zanetti, Ioannis Katsavounidis, Alan C. Bovik IEEE Transactions on Image Processing (TIP) 2024 arxiv / project page / database request link
	Exploring Explainability in Video Action Recognition Avinab Saha, Shashank Gupta, Sravan Kumar Ankireddy, Joydeep Ghosh 3rd Explainable AI for Computer Vision (XAI4CV) Workshop, IEEE/CVF Computer Vision and Pattern Recognition (CVPR)* 2024 arxiv / poster / slides
	HIDRO-VQA: High Dynamic Range Oracle for Video Quality Assessment Shreshth Saini, Avinab Saha, Alan C. Bovik 3rd Workshop on Image/Video/Audio Quality in Computer Vision and Generative AI, IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2024 arxiv / code / slides
	Re-IQA: Unsupervised Learning for Image Quality Assessment in the Wild Avinab Saha, Sandeep Mishra, Alan C. Bovik IEEE/CVF Computer Vision and Pattern Recognition (CVPR) 2023 arxiv / code / video / poster/ slides
	Study of Subjective and Objective Quality Assessment of Mobile Cloud Gaming Videos Avinab Saha, Yu-Chih Chen, Chase Davis, Bo Qiu, Xiaoming Wang, Rahul Gowda, Ioannis Katsavounidis, Alan C. Bovik IEEE Transactions on Image Processing (TIP) 2023 arxiv / project page / database request link / slides
	GAMIVAL: Video Quality Prediction on Mobile Cloud Gaming Content Yu-Chih Chen, Avinab Saha, Chase Davis, Bo Qiu, Xiaoming Wang, Rahul Gowda, Ioannis Katsavounidis, Alan C. Bovik IEEE Signal Processing Letters (SPL) 2023 arxiv / code
	Knowledge Distillation Inspired Fine-tuning of Tucker Decomposed CNNs and Adversarial Robustness Analysis Ranajoy Sadhukhan, Avinab Saha, Jayanta Mukhopadhyay, Amit Patra IEEE International Conference on Image Processing (ICIP) 2020 paper / code
	Fitness Based Layer Rank Selection Algorithm for Accelerating CNNs by CANDECOMP/PARAFAC (CP) Decompositions Avinab Saha, K Sai Ram, Jayanta Mukhopadhyay, Partha Pratim Das, Amit Patra IEEE International Conference on Image Processing (ICIP) 2019 paper

Demos & Invited Talks/ Papers

[Talk] On SIGGRAPH '25 Paper FaceExpressions-70k, VQEG, Spring Meeting, May 2025 [Video]
[Talk] Guest Lecturer for ECE 371Q : Digital Image Processing, UT Austin, Fall 2023.
[Paper] Perceptual Video Quality Assessment: The Journey Continues!, Frontiers in Signal Processing, June 2023
[Talk] Video Quality Assessment of Mobile Cloud Gaming Videos, VQEG, Spring Meeting, June 2023 [Slides]
[Demo] 3D Human Avatars in Virtual Reality, 6G@UT Forum, UT Austin, March 2023 [Media]

Community Involvement

Co-organized the 4th Explainable AI for Computer Vision Workshop at CVPR 2025 [June 2025]
Assistant Director, LIVE [Sept 2023 - April 2025]
Peer Reviewer for CVPR, ICLR, NeurIPS, ICML, IEEE TPAMI
Graduate Student Peer Mentor, ECE, UT Austin [Sept 2021 - Dec 2022]
Workplace Happiness Task Force, Samsung Research [Jan 2020 - Jan 2021]
Under-Graduate Council Member, IIT Kharagpur [July 2018 - May 2019]

Source code credit to Dr. Jon Barron