FishDTecTools : Fish Detection Solution Using Neural Network Approach

Detection and recognizing an object is very difficult and take a long times especially for the live object. In fisheries industries, species of the fish must be recognizing for calculating their population. Currently, manual approach is done by an official of fisheries Department one by one for calculating and measuring fish length. One of the fundamental challenging problems in computer vision is detecting object inside an image or video frames. This paper is presenting a neural network-based approach for detecting fish object inside a digital image. The neural network is used to recognize the fish object from a small window that scan for the object all over the image. The neural network will examine the windows of the image and decides whether each window contains a fish. This approach is aimed to eliminate the difficult task of manually selecting the fishes, which must be chosen and localized to be analyzed. This work is a part of process on detecting the fish, counting and estimating the length of the fish.