An Efficient Human Activity Recognition Technique Based on Deep Learning
- Авторы: Khelalef A.1, Ababsa F.2, Benoudjit N.1
-
Учреждения:
- Laboratoire d’Automatique Avancée et d’Analyse des Systèmes (LAAAS), University Batna-2
- Ecole Nationale d’Arts et Métiers, Institut image-Le2i
- Выпуск: Том 29, № 4 (2019)
- Страницы: 702-715
- Раздел: Applied Problems
- URL: https://journals.rcsi.science/1054-6618/article/view/195751
- DOI: https://doi.org/10.1134/S1054661819040084
- ID: 195751
Цитировать
Аннотация
In this paper, we present a new deep learning-based human activity recognition technique. First, we track and extract human body from each frame of the video stream. Next, we abstract human silhouettes and use them to create binary space-time maps (BSTMs) which summarize human activity within a defined time interval. Finally, we use convolutional neural network (CNN) to extract features from BSTMs and classify the activities. To evaluate our approach, we carried out several tests using three public datasets: Weizmann, Keck Gesture and KTH Database. Experimental results show that our technique outperforms conventional state-of-the-art methods in term of recognition accuracy and provides comparable performance against recent deep learning techniques. It’s simple to implement, requires less computing power, and can be used for multi-subject activity recognition.
Об авторах
A. Khelalef
Laboratoire d’Automatique Avancée et d’Analyse des Systèmes (LAAAS), University Batna-2
Автор, ответственный за переписку.
Email: Khelalef_aziz@yahoo.fr
Алжир, Batna
F. Ababsa
Ecole Nationale d’Arts et Métiers, Institut image-Le2i
Автор, ответственный за переписку.
Email: fakhreddine.ababsa@ensam.eu
Франция, Paris
N. Benoudjit
Laboratoire d’Automatique Avancée et d’Analyse des Systèmes (LAAAS), University Batna-2
Автор, ответственный за переписку.
Email: n.benoudjit@univ-batna2.dz
Алжир, Batna