A New Loss Function for Simultaneous Object Localization and Classification

Robots play a pivotal role in the manufacturing industry. This has led to development of computer vision. Since AlexNet won ILSVRC, convolutional neural networks (CNNs) have achieved state-of-the-art status this area. In work, novel method is proposed simultaneously detect and predict localization objects using custom loop CNN, performing two most important tasks vision with single method. Two different loss functions are evaluate compare results. The obtained results show that network able perform both accurately, classifying images correctly locating precisely. Regarding functions, when target classification values computed, performs better task. Following improvements expected be made task by refining training processes functions.