The number of older people in different countries are constantly increasing. Most of this people prefer to live independently. Falls may lead to serious injuries and may even cause death of people. As a solution to this problem it is essential to develop a fall detection system. The objective of this project is to identify and detect unusual activity for an elderly person. Individuals spend the majority of their time in their home or workplace and many feels that these places are their sanctuaries. The information about the person is stored in a database. So in an emergency situation the neighbor can go through the details of the affected person and he/she can refer all the details about the affected person. A camera is continuously capturing the video of the bedridden person. Machine learning techniques use the information to identify and reason about normal behavior in terms of recognized and forecasted activities. Once the abnormal behavior is identified as a threat, a message is sent to the neighbor or corresponding authorities. In most emergency cases, the elderly patient seek in-patient care, which is very expensive and can be a serious financial burden on the patient if the hospital stay is prolonged, and it won’t be affordable for everyone. The proposed work allows people to remain in their comfortable home environment rather than inexpensive and limited nursing homes or hospitals, ensuring maximum independence to the occupants. Therefore, an affordable and comprehensive healthcare solution with minimal workforce have much importance for longterm health management and population. We make use of Artificial Intelligence, Machine Learning, and computer vision

Keywords : Activities of Daily Living (ADL), Support Vector Machine (SVM), Red Green Blue (RGB), Hidden Markov Models (HMM), Sensor of Movements (SoM), Remote Telecare Center (RTC), Decision and Analysis Device (DAD), Human Activity Recognition (HAR).