About me
I am Professor in the Manning College of Information and Computer Sciences at the University of Massachusetts Amherst, where I also serve as Co-Director of the Computer Vision Lab. Prior to joining UMass Amherst, I was a Research Assistant Professor at TTI Chicago. I obtained my PhD under the supervision of Jitendra Malik from the University of California, Berkeley, in 2011, and a BTech in Computer Science and Engineering from IIT Kanpur in 2006. In the past, I have enjoyed working at the University of Amsterdam, AWS AI, Google, INRIA, Microsoft Research, the CLSP Center at JHU, and Oxford University. My research is funded by the National Science Foundation, including a CAREER award, NASA, and awards from Climate Change AI, Facebook, NVIDIA, Dolby, and Adobe.
My research focuses on high-level visual recognition algorithms. I am also interested in AI applications for science, particularly in ecology and remote sensing. For an overview of my research, take a look at my recent publications, CV, and research statement (visualized as a word cloud). I am also on Google Scholar.
A better computer science department rankings for PhD students: csrankings.org.
News
+ Zezhou will present the LU-NeRF at the main conference.
+ At the R6D workshop, Zezhou and Matheus will present accidental turntables.
+ Rangel and Aaron will present COSE, an explanability benchmark, and Oindrila will present PARTICLE, a self-supervised technique for fine-grained tasks, at the VIPriors workshop.
+ I am also an invited speaker at the VIPriors workshop.
+ I gave a talk at the deep declarative networks workshop.
+ Chenyun and Matheus presented their papers at the main conference.
+ Matheus and Jong-Chyi presented their papers at the visual learning with limited labels workshop.
+ Jong-Chyi co-organized the semi-supervised challenge at the FGVC7 workshop.
+ Talks and competition results from the FGVC7 workshop posted.
+ MistNet appears in Methods in Ecology and Evolution (press release)
+ Two papers to appear at ICCV 19 and one at 3DV 19
+ Two papers to appear at CVPR 19 and one at AAAI 19, UBICOMP 19, KDD 19
+ Journal accepted on IEEE Trans. on Info. Theory
+ ArXiv highlights: Task2Vec, PrGAN++
+ Several competitions are live. Participate to win prizes!
+ Call for papers posted.
+ VisemeNet to appear @ SIGGRAPH 18
+ Two papers: SPLATNet, CSGNet to appear @ CVPR 18
+ Fine-Grained Visual Categorization (FGVC5) workshop @ CVPR 18
+ 2 papers at 3DV, 3 papers at BMVC 17, 1 at ICCV 17
+ Bilinear CNNs PAMI preprint, arXiv v6, Jun 2017
+ 3D Shape Segmentation with Projective Convolutional Networks, CVPR 17 (oral)
+ 3D Shape Reconstruction from Sketches via Multi-view Convolutional Networks, arXiv, July 17
Recent research highlights
- Active Measurement: Efficient Estimation at Scale, Hamilton et al., NeurIPS 2025
- Generate, Transduct, Adapt: Iterative Transduction with VLMs, Saha et al., ICCV 2025
- WildSAT: Learning Satellite Image Representations from Wildlife Observations, Daroya et al., ICCV 2025
- Combining Observational Data and Language for Species Range Estimation, Hamilton et al., NeurIPS 2024
- Task2Box: Box Embeddings for Modeling Asymmetric Task Relationships, Daroya et al., CVPR 2024
- Using Spatio-Temporal Information in Weather Radar Data to Detect and Track Communal Bird Roosts, Perez et al., Remote Sensing in Ecology and Conservation, 2024
- Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions, Saha et al., CVPR 2024
- DISCount: Counting in Large Image Collections with Detector-based Importance Sampling, Perez et al., AAAI 2024. Best Paper for the AI for Social Impact Track