We present another PC vision based framework for powerful activity sign acknowledgment and following. Such a framework displays an indispensable help for driver help with a canny car. Right off the bat, a shading based division strategy is connected to create movement sign hopeful areas. Also, the HoG highlights are separated to encode the identified activity signs and after that producing the element vector. This vector is utilized as a contribution to a SVM classifier to distinguish the activity sign class. At last, a following strategy in view of optical stream is performed to guarantee a ceaseless catch of the perceived activity sign while quickening the execution time. Our strategy bears high accuracy rates under various testing conditions.