Vision and AI Laboratory (VAL) -- IISc

SFDA-Seg

Generalize then Adapt: Source-free Domain Adaptation for Semantic Segmentation (ICCV 2021)

GAMA-GAT

Guided Adversarial Attack for Evaluating and Enhancing Adversarial Defenses, NeurIPS Spotlight 2020

css-ccnn

Implementation for "Completely Self-Supervised Crowd Counting via Distribution Matching" (http://arxiv.org/abs/2009.06420)

ss_human_mesh

Code repository for the paper: Appearance Consensus Driven Self-Supervised Human Mesh Recovery

USFDA

Universal Source-Free Domain Adaptation (CVPR 2020)

InheriTune

Code Release for the CVPR 2020 (oral) paper, "Towards Inheritable Models for Open-set Domain Adaptation".

BPFC

Towards Achieving Adversarial Robustness by Enforcing Feature Consistency Across Bit Planes

ssl_3d_recon

Code release for "From Image Collections to Point Clouds with Self-supervised Shape and Pose Networks" (CVPR 2020)

fast-feature-fool

Data independent universal adversarial perturbations

SAT-Rx

Regularizers for Single-step Adversarial training

pose_estimation

Code for our work on pose-estimation using template 3D models.

GANTree

Code for ICCV Paper "GAN-Tree: An Incrementally Learned Hierarchical Generative Framework for Multi-Modal Data Distributions"

DS-263-VA

crowd-counting-scnn

This project is an implementation of the crowd counting model proposed in our CVPR 2017 paper - Switching Convolutional Neural Network(SCNN) for Crowd Counting. SCNN is an adaptation of the fully-convolutional neural network and uses an expert CNN that chooses the best crowd density CNN regressor for parts of the scene from a bag of regressors. This helps it tackle intra-scene crowd density variation and obtain SOTA results

differ

Codes release for "DIFFER: Moving Beyond 3D Reconstruction with Differentiable Feature Rendering", (CVPRW '19)

capnet

Code release for "CAPNet: Continuous Approximation Projection For 3D Point Cloud Reconstruction Using 2D Supervision", (AAAI-19)

3d-psrnet

Repository for 3D-PSRNet: Part Segmented 3D Point Cloud Reconstruction [ECCVW 2018]

3d-lmnet

Repository for 3D-LMNet: Latent Embedding Matching for Accurate and Diverse 3D Point Cloud Reconstruction from a Single Image [BMVC 2018]

Pose2vec

A Repository for maintaining various human skeleton preprocessing steps in numpy and tensorflow along with tensorflow model to learn pose embeddings.

densepcr

Repository for 'Dense 3D Point Cloud Reconstruction Using a Deep Pyramid Network' [WACV 2019]

DLCV

Webpage for DS265: Deep Learning in Computer Vision Course

DLCV2018

Webpage for DS265: Deep Learning in Computer Vision Course

GD-UAP

Generalized Data-free Universal Adversarial Perturbations

sketchguess

Repository for code, models and datasets for the paper Game of Sketches: Deep Recurrent Models of Pictionary-style Word Guessing accepted at 36th AAAI Conference on Artificial Intelligence (AAAI-18), New Orleans, USA

gat

Repository for the Gray-box Adversarial Training project presented at ECCV 2018.

AAA

cnn-fixations

Visualising predictions of deep neural networks

nag

[CVPR 2018] Tensorflow implementation of NAG : Network for Adversary Generation

DLCV-assignments

Repository containing assignments for DS-265 Deep Learning for Computer Vision

deligan

This project is an implementation of the Generative Adversarial Network proposed in our CVPR 2017 paper - DeLiGAN : Generative Adversarial Networks for Diverse and Limited Data. DeLiGAN is a simple but effective modification of the GAN framework and aims to improve performance on datasets which are diverse yet small in size.