Abnormal Gait Recognition Based on RBF Neural Network

Gait, the manner of walking, is a very common behavior that we can observe in daily life. Analyzing human gait has stimulated considerable interests in recent machine vision researches. So far, contributions to this topic exclusively are dealt with the tasks of person identification and activity recognition. Gait analysis has not been used in the aid of home care system or other surveillance application to detect abnormal cases since analysis is only done on normal types of walking. In this paper, we propose a simple but effective approach to abnormal gait analysis using computer vision technique. After segmenting the foreground object, we extract the feature vector reflecting the abnormal gait from the aspect ratio and achieve the abnormal gait recognition by RBF network. Experimental results show its promising performance.