RetinaNet Based Environment Classification
Ridhanya Sree Balamurugan, R. S. Arunkumar, S. Mohan · Asian Journal of Computer Science and Technology · 2018
Environmental classification is very useful for visually impaired persons and Robotic applications. The main objective of this work is to detect and recognize the objects present in a scene and identify the environment based on the occurrence probability of the objects in the scene. Objects from the real-time images are detected and recognized by means of RetinaNet. Occurrence probabilities of the recognized objects are used to identify the environment.