Abhisek Ray

Shapley-Neuron-Values. [ICML 2026] Official code of SNV paper.

lightly-train. All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.

only-what-s-necessary. Python

From-Pixels-to-Privacy-Temporally-Consistent-Video-Anonymization-via-Token-Pruning-for-Privacy-Pres. Python

Balancing-Privacy-and-Action-Performance-A-Penalty-Driven-Approach-to-Image-Anonymization. Official implementation of the CVPRW '25 paper "Balancing Privacy and Action Performance: A Penalty Driven Approach to Image Anonymization"-- Accepted for CVPRW

Saliency_Map_in_DL. Python

PATReId-Pose-Apprise-Transformer-Network-for-Vehicle-Re-Identification. Python

fastdup. fastdup is a powerful, free tool designed to rapidly generate valuable insights from image and video datasets. It helps enhance the quality of both images and labels, while significantly reducing data operation costs, all with unmatched scalability.

fiftyone. Refine high-quality datasets and visual AI models

ViFM_Survey. Foundation Models for Video Understanding: A Survey

CVPR-2024-Papers.