Review on Recent Advances in Human Action Recognition in Video Data

Human action recognition has been applied in many fields, such as video surveillance and human computer interaction, where it helps to improve performance. Numerous reviews of the literature have been done, but rarely have these reviews concentrated on skeleton-graph-based approaches. Connecting the skeleton joints as in the physical appearance can naturally generate a graph. This paper provides an up-to-date review for readers on skeleton graph-neural-network-based human action recognition. After analyzing previous related studies, a new taxonomy for skeleton-GNN-based methods is proposed according to their designs, and their merits and demerits are analyzed. In addition, the datasets and codes are discussed. Finally, future research directions are suggested.

show abstract

“…Papers on modalities [ 10 , 19 , 20 , 21 , 22 , 23 ] on videos (images) [ 24 , 25 , 26 ] on skeletons, and [ 27 ] analyzed multi-modalities.…”

Section: Previous Workmentioning

confidence: 99%

Skeleton Graph-Neural-Network-Based Human Action Recognition: A Survey

Feng

Meunier

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Baisware A. et al [10] focused on the most searched study area, the video-based human activity recognition since it has so many potential uses. Deep learning-based human activity recognition is one of the streams that is constantly developing and looking promising with the steady rise of high-performance computers, computer vision, and GPUs.…”

Section: Related Workmentioning

confidence: 99%

Classification of Videos based on Object Detection in Video Files

Srin

Baydeti

2022

Preprint

View full text Add to dashboard Cite

The primary objective of this work is to classify the videos, which comprises the extraction of video frames and generation of features vector and identification of action sequence in the video. A deep neural network has been proposed that combines Convolutional Neural Network and Long Short Term Memory models to achieve better spatial and temporal features. Video Classification is performed by extracting the features from the frames by considering the parameters viz., frame width and height, video sequence length etc. In the proposed model, the frames extracted from a video are fed into a two-layer LSTM. The outcome of the LSTM model is forwarded to the Convolution Layers with an additional Global Average Pooling (GAP) layer in place of the fully connected layer. The frames extracted from the video are sent as an input to the twolayered Convolutional 2D layer, which is followed by Batch normalization and Max Pooling. Fully connected layer is replaced by Global Average Pooling (GAP) layer followed by dense layer. Depending on the input data, more number of batch normalization and dense layers would be added in order to achieve more accuracy. UCF101 Dataset is considered to classify the videos. The results demonstrate that the LRCN methodology outperforms both the conventional CNN method and ConvLSTM in terms of prediction accuracy. The suggested approach also provides more accurate temporal and spatial stream identification. The results have shown that the probability of the action sequence is predicted to be around 82 percentage with an accuracy of 92.86 percentage.

show abstract

“…After the release of the Kinect v1 and Kinect v2, several gait databases [3,4,6,14,33,37] have been collected to support model-based gait recognition. They consider various factors such as speed [4,33], carrying [4,14,33,37], view [6,14] except occlusion caused by clothing and multi-gait, which are very common in practice.…”

Section: Gait Recognition Databasesmentioning

confidence: 99%

A Benchmark for Gait Recognition under Occlusion Collected by Multi-Kinect SDAS

Li¹,

Zhao²

2021

Preprint

View full text Add to dashboard Cite

Human gait is one of important biometric characteristics for human identification at a distance. In practice, occlusion usually occurs and seriously affects accuracy of gait recognition. However, there is no available database to support in-depth research of this problem, and state-of-arts gait recognition methods have not paid enough attention to it, thus this paper focuses on gait recognition under occlusion.We collect a new gait recognition database called OG RGB+D database, which breaks through the limitation of other gait databases and includes multimodal gait data of various occlusions (self-occlusion, active occlusion, and passive occlusion) by our multiple synchronous Azure Kinect DK sensors data acquisition system (multi-Kinect SDAS) that can be also applied in security situations. Because Azure Kinect DK can simultaneously collect multimodal data to support different types of gait recognition algorithms, especially enables us to effectively obtain camera-centric multi-person 3D poses, and multi-view is better to deal with occlusion than single-view. In particular, the OG RGB+D database provides accurate silhouettes and the optimized human 3D joints data (OJ) by fusing data collected by multi-Kinects which are more accurate in human pose representation under occlusion. We also use the OJ data to train an advanced 3D multi-person pose estimation model to improve its accuracy of pose estimation under occlusion for universality. Besides, as human pose is less sensitive to occlusion than human appearance, we propose a novel gait recognition method SkeletonGait based on human dual skeleton model using a framework of siamese spatio-temporal graph convolutional networks (siamese ST-GCN). The evaluation results demonstrate that SkeletonGait has competitive performance compared with state-of-art gait recognition methods on OG RGB+D database and popular CAISA-B database. This database will be available at https://github.com/cvNXE/OG-RGB-D-gait-database.

show abstract

Review on Recent Advances in Human Action Recognition in Video Data

Cited by 12 publications

References 7 publications

Skeleton Graph-Neural-Network-Based Human Action Recognition: A Survey

Skeleton Graph-Neural-Network-Based Human Action Recognition: A Survey

Classification of Videos based on Object Detection in Video Files

A Benchmark for Gait Recognition under Occlusion Collected by Multi-Kinect SDAS

Contact Info

Product

Resources

About