Intelligent Intrusion Detection Using TensorFlow

Detection of human beings accurately in a visual system like in images, videos, live stream videos is critical for diverse application which includes elderly abnormality detection, violent detection, human gait characterization, congestion analysis, person recognition and gender classification. Object detection is the process of distinguishing the presence of predefined types of objects in a frame. This process includes both identification of the presence of the objects with a rectangular bounding surrounding it. The application utilizes MSCOCO model trained from ImageNet database with TensorFlow to detect human beings or objects with a greater accuracy. For greater speed in detection, single shot detectors (SSD) have been used for detection. There is a tradeoff between faster detection and accuracy. The test observations results show the detection as well as the recognition of the human beings with a greater speed and accuracy that indicates the effectiveness of the proposed approach.