Thin-plate spline interpolation estimates scene illumination for better color accuracy in images. This method enhances automatic white balancing for digital cameras by using a refined training set. 🛠️🌈

[NeurIPS 2023] Official implementation of the paper "CAST: Cross-Attention in Space and Time for Video Action Recognition"

This repository contains my personal code for the paper Learning Spatiotemporal Features with 3D Convolutional Networks by Du Tran, Lubomir Bourdev, Rob Fergus, Lorenzo Torresani, Manohar Paluri.

SoccerAct10 is a dataset which contains 10 different soccer actions. This dataset was developed using the videos from YouTube.

A PyTorch implementation of RSC based on MMM 2023 paper "Weakly-supervised Temporal Action Localization with Regional Similarity Consistency"

NTU Fall 2019 Deep Learning for Computer Vision Homework 4 - Action Segmentation

A PyTorch implementation of MCA based on PRICAI 2022 paper "Weakly-supervised Temporal Action Localization with Multi-head Cross-modal Attention"

Official code for CVPR'19 paper "Collaborative Spatiotemporal Feature Learning for Video Action Recognition"

A simple but efficient transformer model for video action recognition

A PyTorch implementation of UM based on AAAI 2021 paper "Weakly-supervised Temporal Action Localization by Uncertainty Modeling"

Code associated with "M2A: Motion Aware Attention for Accurate Video Action Recognition"

This is my todo list and some useful materials

Single run action detector on video stream data

Video Action Recognition using Deep Learning

[Arxiv2020] The code for our paper 《Self-Supervised Temporal-Discriminative Representation Learning for Video Action Recognition》 https://arxiv.org/abs/2008.02129

This project involves the identification of different actions from video clips where the action may or may not be performed throughout the entire duration of the video. This is done using two CNN models which are 3D-CNN and LSTM models.

A new and computationally cheap method to perform human activity recognition using PoseNet and LSTM. Where we use PoseNet for Preprocessing and LSTM for understand the sequence.

3D backbone for video action recognition

An Online Web Game "You Perform, I Guess!" based on C3D Model

MFNet的实现和改进。