activitynet

Awesome video understanding toolkits based on PaddlePaddle. It supports video data annotation tools, lightweight RGB and skeleton based action recognition model, practical applications for video tagging and sport action detection.

An official implementation for "CLIP4Clip: An Empirical Study of CLIP for End to End Video Clip Retrieval"

[CVPR 2021] Actor-Context-Actor Relation Network for Spatio-temporal Action Localization

An official implementation for "X-CLIP: End-to-End Multi-grained Contrastive Learning for Video-Text Retrieval"

An VideoQA dataset based on the videos from ActivityNet

Adversarial Background-Aware Loss for Weakly-supervised Temporal Activity Localization (ECCV 2020)

Salient Video Frames Sampling Method Using the Mean of Deep Features for Efficient Model Training (KIBME 2021)

Video classification in TensorFlow using Mask R-CNN. This project is built upon https://github.com/matterport/Mask_RCNN. The dataset used to train Mask R-CNN has been built with LabelBox, the video classification has been done with an LSTM that classifies activities taken from a subset of ActivityNet dataset (Gymnastics activities). This repository has been used for the final project of the module Vision and Perception (Spring 2018) at Sapienza University of Rome.

An Activity-Net Downloader,written in Java.