Video behavior recognition algorithm based on two-stream heterogeneous convolutional neural network
نویسندگان
چکیده
Abstract In video behavior recognition, making full use of the spatio-temporal information contained in frame is critical point to further improve recognition accuracy. our study, we propose a algorithm based on two-stream heterogeneous network, so that can extract feature frame, and finally take advantage features sequence. view characteristics RGB optical flow images, this paper uses DenseNet121 Inception-V4 structure fully video. paper, UCF101 dataset used evaluate algorithm, accuracy 91.7%, which verified reliability proposed algorithm.
منابع مشابه
EMG-based wrist gesture recognition using a convolutional neural network
Background: Deep learning has revolutionized artificial intelligence and has transformed many fields. It allows processing high-dimensional data (such as signals or images) without the need for feature engineering. The aim of this research is to develop a deep learning-based system to decode motor intent from electromyogram (EMG) signals. Methods: A myoelectric system based on convolutional ne...
متن کاملVideo Based Face Recognition Using Convolutional Neural Network
This chapter addresses an improved approach to video face recognition (VFR). Techniques to recognize faces in video streams have been described in the literature for more than 20 years (Wang et al., 2009). Early methods were based on the still-to-still techniques which aimed at selecting good frame and did some relative processing. Recently researchers began to truly solve such problems by spat...
متن کاملTwo-Stream 3D Convolutional Neural Network for Skeleton-Based Action Recognition
It remains a challenge to efficiently extract spatialtemporal information from skeleton sequences for 3D human action recognition. Although most recent action recognition methods are based on Recurrent Neural Networks which present outstanding performance, one of the shortcomings of these methods is the tendency to overemphasize the temporal information. Since 3D convolutional neural network(3D...
متن کاملTwo-Stream convolutional nets for action recognition in untrimmed video
We extend the two-stream convolutional net architecture developed by Simonyan for action recognition in untrimmed video clips. The main challenges of this project are first replicating the results of Simonyan et al, and then extending the pipeline to apply it to much longer video clips in which no actions of interest are taking place most of the time. We explore aspects of the performance of th...
متن کاملConvolutional Neural Network-based Place Recognition
Recently Convolutional Neural Networks (CNNs) have been shown to achieve state-of-the-art performance on various classification tasks. In this paper, we present for the first time a place recognition technique based on CNN models, by combining the powerful features learnt by CNNs with a spatial and sequential filter. Applying the system to a 70 km benchmark place recognition dataset we achieve ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Journal of physics
سال: 2022
ISSN: ['0022-3700', '1747-3721', '0368-3508', '1747-3713']
DOI: https://doi.org/10.1088/1742-6596/2258/1/012028