Abstract
Suspicious activity to predict a body part or joint of a person from a photo or video. This project will involve the discovery of a suspicious person Work from real-time CCTV footage using neural networks. Suspicious human activity is one of the major problems in the computer industry studied for over 15 years. It is important because of the small value of applications that can benefit from Finding a Job. For example, human posture measurements are used in programs that include video surveillance, animal tracking and behavior comprehension, advanced sign language detection. Human computer interaction, and small motion capture. Low cost Deep sensors have limited limitations in home use, and their low adjustment and knowledge of sound depth make it difficult to measure a person's posture. From the depths of the images. Therefore, we plan to use neural networks to overcome these problems. Suspicious recognition of human activity from surveillance video is an effective research site for image processing and computer perspective. Through the visible observation, human activities can be recognized in a sensitive and public environment places like bus stations, train stations, airports, banks, shopping malls, schools and colleges, parking lots, roads, etc. prevention of terrorism, theft, accidents and illegal parking, vandalism, fighting, chain-taking, crime and other suspicious activities. It is very difficult to watch public places on an ongoing basis, so it requires a smart video surveillance that can be monitored human activities in real time and classify them as normal and unusual jobs; and can create a warning. In particular, in the research conducted opt-out is in photos and not videos. Also, there are no published papers attempting to use CNN for suspicious activity.