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
