HUMAN ACTIVITY DETECTION BASED ON EDGE POINT MOVEMENTS AND SPATIO-TEMPORAL FEATURES IN INDOOR

As the population of elder people increasing rapidly, the need for the health care systems is also increasing. To provide a better health care and improving the quality of living of elder/disabled persons who resides independently, ‘Activity Recognition’ systems can used. The daily activities of the persons and the abnormal activities can be detected using the activity recognition systems. There are s o many methods exist today for this purpose. These nsor based methods and video data based systems are the existing techniques. Often these systems are complex and not guaranteed the 100% activity detection rate. To overcome the cost and complexity of the existing methods, here proposes a new system with simple mechanisms and guaranteed a better performance. The input to the system is the visual data, taken from multiple cameras placed in the home/indoor.