Detection of human being and non-human object from image and video sequences

This paper represents a new method to classify human or non-human objects from the image and video sequences. A clear detection is essential for such kind of classification. The difficulties that arise during the classification of human being are clothing, illuminations, poses, shadows and shape variations. If the quality of video is not good enough then detection cannot be done perfectly. To avoid these difficulties, we use FCM and k-means based new thresholding technique to decide the threshold point for object detection. Detected object is human being or not is an important task in case of space, security purpose and visual surveillance. So, after the successful detection of object, we use golden ratio as an object classifier to classify it as human or non-human. It is a ratio related to human body parts, such as, height and width of a human body. This ratio also called as divine proportion.