Human Action Recognition Using Squeezed Convolutional Neural Network
Hakim Nasaoui, Insaf Bellamine, Hassan Silkan · 2022
In recent years, Convolutional Neural Network (CNN) gained great success in many applications, especially in computer vision. In this paper we propose a novel deep learning approach for Human Action Recognition inspired by one of the most popular architectures being SqueezeNet with an incredibly small model size. The results show that the proposed model performs well on a KTH dataset as well as UCF-Sport dataset