Abstract
Human activity recognition is an important task in computer vision because it has many application areas such as, healthcare, security, entertainment, and tactical scenarios. This paper presents a methodology to automatically recognize human activity from input video stream using Histogram of Oriented Gradient Pattern History (HOGPH) features and SVM classifier. For this purpose, the proposed system extracts HOG features from a sequence of consecutive video frames and analyzes them to construct HOGPH feature vector. The HOGPH feature vectors are used to train a multi-class SVM classifier for different human activities. In test mode, we use the classifier with HOGPH feature vector to recognize human activity. We have experimented with video data of human activity in real environments for three different tasks (browsing, reading, and writing). The experimental result and its accuracy reveal that the proposed system is applicable to recognize human activity in real-life.
Highlights
Recognizing human activity or task from real-time video data is one of the promising and challenging applications of computer vision
We recognize human task or acitivity in unconstrained real-world environments, such as Internet browsing in an office environment, reading and writing in an library environment etc., using Histogram of Oriented Gradient Pattern History (HOGPH) and Support Vector Machine (SVM) classifier
The proposed activity recognition system first extracts histogram of oriented gradient features (HOG) features of activity during a certain period of time to build up an activity pattern known as the histogram of oriented gradient pattern history (HOGPH)
Summary
Recognizing human activity or task from real-time video data is one of the promising and challenging applications of computer vision. In this research, we proposed an activity recognition system that tries to assess the qualitative characteristics of the human activity For this purpose, the proposed activity recognition system first extracts HOG features of activity during a certain period of time to build up an activity pattern known as the histogram of oriented gradient pattern history (HOGPH). To develop a good human activity recognition system, we need to develop a pattern recognition system that is robust to intra-class variability. Intra-class variability can happen if an activity is performed by the same person at different time. For activity recognition, we need a classifier that adapts this type of intra-class variability To address this issue, the proposed system uses the multi-class SVM classifier with an increase amount of training data for different activity classes using liblinear kernel. The proposed HOGPH features are highly discriminative and person independent that make help the system to adapt intra-class variability
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have