Gesture Recognition System to Identify Animal Gestures

Gesture recognition is a technique that lets us control any electronic device by identifying the movements or positions of different human body parts. Human gestures are a type of non-verbal interaction that can represent the most instinctive and unique approach for communicating with computers. When compared with all possible instances for gesture technology now available, there is no method for interpreting animal gestures. The objective of our project is to create an animal gesture recognition system that recognizes emotions of animals such as dogs and cats. The animal posture, such as the tail’s position, hand movements and eye movements will be initially recovered and subsequently used for model training. A customized CNN method is used in the proposed system to classify the datasets. Currently there is no dataset available for validating the system. The user provides an image of the animal as input into the detector system. A dataset is created which contain images collected from the web is trained and processed and then comparison of input images is done. The corresponding classification of emotions such as anger, fear, happiness are conveyed by the gestures provided into the system displayed as the output. To get accurate results, features of images are used and after applying successful methodology for detection and recognition of the image, the system has obtained 95.6% accuracy using practical and user-friendly models and displayed as visuals