Abstract
Classification of objects in visual-images is a hot open research problem in the area of computer vision, especially in modern Internet of Things (IoT) systems. In this paper,a robust and reliable object taxonomy purposed framework is presented. The proposed model functionalize the research future targeted, deep convolutional neural network components to construct a Robust Object Recognition Network (RORNet). The RORNet consists of four convolutional layers, four Relu/Leaky Relu activation layers, three max-pooling layers and only two fully connected layers for extracting expected input image features. To speedup the training process, we used non-saturated neurons with a very efficient Graphics Processing Unit (GPU) coding for the convolution operation. To minimize the over-fitting issue in the fully-connected layers, we fictionalize a recently-developed regularization approach "dropout" with a dropping probability of 55%. The experimental and simulations results show that proposed RORNet framework has a high potential capability in the recognition of unseen images.