Anshul Shah

I am a Research Scientist at Apple. I graduated with a Ph.D. in Computer Science from Johns Hopkins University in Feb'24 where I was advised by Prof. Rama Chellappa and was working on problems in Computer Vision and Machine Learning. I also obtained an MS in Computer Science from University of Maryland, College Park.

Previously I obtained a B.Tech(Honors) and M.Tech in Electrical Engineering from Indian Institute of Technology Madras. At IIT-M I was working under the guidance of Prof. A.N Rajagopalan in the area of Image Deblurring and Video Generation.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github

profile photo

  • 03/2024: Started as a Research Scientist at Apple
  • 02/2024: Successfully defended my PhD dissertation
  • 10/2023: Invited Talk on STEPs at ICCV's CV4Metaverse workshop
  • 10/2023: Proposal on SSL co-written with PI Rama Chellappa was awarded Amazon Research award (2023-24).
  • 07/2023: Paper accepted at ICCV 2023
  • 03/2023: Started Research Internship at Apple's ML Research org with Raviteja, Gierad, Anurag and Karren
  • 10/2022: Named as an Amazon Fellow as a part of JHU + Amazon initiative for Interactive AI! [AI2AI,Amazon Science]


I am interested in computer vision, machine learning. Specifically, my current research interests are in Self-Supervised learning, learning from complex videos and learning from multiple modalities and sensors.

STEPs STEPs: Self-Supervised Key Step Extraction from Unlabeled Procedural Videos

Anshul Shah, Benjamin Lundell, Harpreet Sawhney, Rama Chellappa
ICCV 2023

We address the problem of extracting key steps from unlabeled procedural videos

HaLP HaLP: Hallucinating Latent Positives for Skeleton-based Self-Supervised Learning of Actions

Anshul Shah, Aniket Roy*, Ketul Shah*, Shlok Mishra, David Jacobs, Anoop Cherian, Rama Chellappa
CVPR 2023

We hallucinate latent postives for learning skeleton encoders without labels

DiffAlign DiffAlign : Few-shot learning using diffusion based synthesis and alignment

Aniket Roy, Anshul Shah*, Ketul Shah*, Anirban Roy, Rama Chellappa
Under Submission

We leverage the recent success of the generative models for few-shot learning

MVA Multi-View Action Recognition using Contrastive Learning

Ketul Shah, Anshul Shah, Chun Pong Lau, Celso M. de Melo, Rama Chellappa
WACV 2023

A method for RGB-based action recognition using multi-view videos.

ObjAwareCrop Object-Aware Cropping for Self-Supervised Learning
Shlok Mishra, Anshul Shah, Ankan Bansal, Abhyuday Jagannatha, Abhishek Sharma, David Jacobs, Dilip Krishnan
TMLR 2022

A novel cropping strategy for SSL on uncurated datasets

FeLMI FeLMi : Few shot Learning with hard Mixup

Aniket Roy, Anshul Shah, Ketul Shah, Prithviraj Dhar, Anoop Cherian, Rama Chellappa
NeurIPS 2022

We propose hard mixup to improve few shot learning

AnisotropicImageNet Learning Visual Representations for Transfer Learning by Suppressing Texture

Shlok Mishra, Anshul Shah, Ankan Bansal, Abhyuday Jagannatha, Jonghyun Choi, Abhinav Shrivastava, Abhishek Sharma, David Jacobs
BMVC 2022

Anisotropic diffusion based augmentation to reduce texture bias in supervised and self-supervised approaches

MMCL Max-Margin Contrastive Learning

Anshul Shah*, Suvrit Sra, Rama Chellappa, Anoop Cherian*
AAAI 2022

We propose a novel objective for contrastive learning motivated by SVMs

arXiv / video / code
PoseAction Pose and Joint-Aware Action Recognition

Anshul Shah, Shlok Mishra, Ankan Bansal, Jun-Cheng Chen, Rama Chellappa, Abhinav Shrivastava WACV 2022

We present a new model and loss for Pose-based action recognition

arXiv / video / code
BringingAlive Bringing Alive Blurred Moments

Kuldeep Purohit, Anshul Shah, A.N. Rajagopalan
CVPR 2019 (Oral)

We present a solution for the goal of extracting a video from a single motion blurred image

AttentionVehicle Attention Driven Vehicle Re-identification and Unsupervised Anomaly Detection for Traffic Understanding

Pirazh Khorramshahi, Neehar Peri, Amit Kumar, Anshul Shah, Rama Chellappa
CVPRW 2019 Nvidia AI City Challenge

An approach for anomaly detection and vehicle Re-ID on the challenging real-world dataset

BlurDetection Learning Based Single Image Blur Detection and Segmentation
Kuldeep Purohit, Anshul Shah, A.N. Rajagopalan
ICIP 2018

A new solution to the problem of obtaining a blur-based segmentation map from a single image affected by motion or defocus blur

Template Credits : Jon Barron