Animal Behavior Detection Using Deep Learning

Abstract

In this competitive modern era, machine learning has been retained the crucial contribution to detect and recognize the object from real-time image data and much more. So one of the most fabulous tasks is ABD, ABD (Animal Behavior Detection) will automate the analysis and recognitions the emotions of an animal such as cat behaviors as monitoring the unobstructed natural environmental image DataSet. In this experiment at firstly, we proposed all technics in the theoretical background. Secondly, we collected entire informatory data such as image data to detect the behavior of animal which illustrates the diverse types of emotions of the cat behavior. In this step, OpenCV could be processing and analysis the images. Thirdly, we used TensorFlow and attempted to fabricate a neural network with different images of cat emotions, and then we attempted to search the correct classification through plenty of test images. We used a deep learning CNN based model (Inception-ResNet-v2) that is used to train all the images from the ImageNet database and classification, also it used to recognize the emotions in static images. Finally, finishing all the required tasks of this application will work thoroughly and will give us nearly 84.5% accuracy