A Lightweight Two-stream Fusion Deep Neural Network Based on ResNet Model for Sports Motion Image Recognition

Niu, Zhiwei

doi:10.1007/s11220-021-00350-6

Cited by 7 publications

(4 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A strong magnetic needle helps to distinguish between vigorous exercise and light exercise [8] . The general step of human motion classification is to preprocess the data after obtaining sensor information, extract features (time-domain features and frequency-domain features) after preprocessing, and design algorithms using the extracted features and filtered data to finally achieve the purpose of motion classification [9] .…”

Section: Data Preparation 21 Data Acquisitionmentioning

confidence: 99%

Human motion recognition of active contrast coding model based on self-attention

Zhou,

Wang,

2023

Second International Conference on Electronic Information Technology (EIT 2023)

View full text Add to dashboard Cite

The demand for high-precision human motion recognition continues to rise. In this experiment, an active contrast coding model based on self-attention is constructed. In the self attention based active contrast coding model, the model fully takes into account the feature differences of different properties of actions, and uses a fixed sliding window to segment the human activity data of inertial sensors with different properties. Cross-validation is used to improve the generalization ability of the model and prevent the occurrence of overfitting. The generalization ability and overfitting phenomenon of the model are evaluated by using the confusion matrix index, ROC curve when each movement is positive and visualization of the model. The accuracy of the final model is 0.987, the F1 score is 0.998, and the precision is 0.987; the recall rate was 0.976, and the AUC values of ROC curves with positive movements were greater than 0.97. Through the visualization of the model, it is found that the boundary between the A3 motion and A8 motion is not clear enough, and the boundary between A18 and A15 motion is not clear enough. Perhaps increasing the amount of data will lead to a more accurate classification of sports.

show abstract

Section: Data Preparation 21 Data Acquisitionmentioning

confidence: 99%

Human motion recognition of active contrast coding model based on self-attention

Zhou,

Wang,

2023

Second International Conference on Electronic Information Technology (EIT 2023)

View full text Add to dashboard Cite

show abstract

“…(15) (16) (17) represents the population information entropy of the algorithm at the t-th iteration;…”

Section: Pheromone Diffusion Modelmentioning

confidence: 99%

“…However, the existing public sports facilities in my country are still only Focusing on providing sports venues, while ignoring the needs of citizens in terms of cultural leisure, social entertainment, etc., there is a lag in the design mode, so how to optimize the design process of such buildings, enhance the cultural and leisure attributes of the main functions, and The fun and experiential nature of key spaces is an issue that needs to be studied urgently [16] . Through the research on the theory of algorithm, a design strategy for the collocation of community sports facilities under the national fitness environment, so as to obtain an outdoor community sports fitness venue suitable for mass sports fitness [17] .…”

mentioning

confidence: 99%

Matching with community sports facilities in a national fitness environment based on ant colony algorithm

Guo

Tian

2023

Preprint

View full text Add to dashboard Cite

The advancement of cultural awareness, the pressure of work is getting higher and higher. This living state has buried many hidden dangers to people's physical and mental health. These "hidden dangers" and various diseases caused by modern lifestyles have made modern physical and mental health, and their enthusiasm for participating in sports is also rising. Under the background of national fitness, all localities have actively responded to the call of the state to build civic sports facilities that serve the masses. However, with the advancement of my country's economic level, the national fitness industry has entered a new era. People are no longer satisfied with having an activity venue, but have an all-round experience of social, leisure, sports and other functions and spaces. become a common pursuit of citizens. Outdoor community sports facilities emerge as the times require, and various sports fitness equipment, sports venues, entertainment and health integration facilities are constantly emerging, forming a modern national fitness sports pattern with community sports as the main feature. Based on the above problems ant colony algorithm theory, this paper proposes a design strategy for the coordination of community sports facilities in a national fitness environment, so as to obtain an outdoor community sports fitness venue suitable for mass physical fitness.

show abstract

“…It is challenging due to many actions, varying camera angles, similarities between actions, and changes in environmental conditions. It has applications in various industries, such as surveillance [2], healthcare [3], eldercare [4], sports [1,5], entertainment [6], and beyond [7].…”

Section: Introductionmentioning

confidence: 99%

Action Recognition in Videos Through a Transfer Learning Based Technique

López-Lozada,

Sossa,

Rubio-Espino

et al. 2024

Preprint

View full text Add to dashboard Cite

In computer vision, human action recognition is a hot topic, popularized by the development of deep learning. Current models have achieved high accuracy results on public datasets. Despite this success, they require significant computational resources for training. Given that transfer learning based techniques allow reusing what other models have already learned and training models with less computational resources, in this work we propose using a transfer learning based approach for action recognition in videos. We describe a methodology for human action recognition using transfer learning techniques in a custom dataset. The proposed method consists of four stages: 1) human detection and tracking, 2) video preprocessing, 3) feature extraction (using pretrained models with ImageNet), and 4) action recognition using a two-stream model consisting of TCNs, LSTMs, and CNNs layers. The custom dataset is imbalanced with 189, 390, 490, 854, and 890 videos per class, respectively. For feature extraction, we analyzed the performance of seven pretrained models: Inception-v3, MobileNet-v2, MobileNet-v3-L, VGG-16, VGG-19, Xception, and ConvNeXt-L. We show that the best results were obtained with the last one. Finally, using pretrained models for feature extraction allowed training in a PC with a single GPU with an accuracy of 94.9%.

show abstract

A Lightweight Two-stream Fusion Deep Neural Network Based on ResNet Model for Sports Motion Image Recognition

Cited by 7 publications

References 28 publications

Human motion recognition of active contrast coding model based on self-attention

Human motion recognition of active contrast coding model based on self-attention

Matching with community sports facilities in a national fitness environment based on ant colony algorithm

Action Recognition in Videos Through a Transfer Learning Based Technique

Contact Info

Product

Resources

About