In this paper we introduce a new methodology for preprocessing images for deep learning classification purposes. We implement Einstein’s photon’s energy law as a preprocessing step before feeding images into a deep learning network. The network is designed mainly with stacked convolutional layers. Optimum architecture of the model and its regularization is tuned experimentally. We use CIFAR10 image dataset for network performance evaluation. We consider each pixel value proportion to the energy of that pixel, considering Red, Green, and Blue (RGB) wavelengths. Results indicate that preprocessing images with Einstein photon’s energy law has significant impact on accuracy of the network in image classification.

This article introduces and explains geometric derivative filter as a new method for image processing and enhancing quality of images.. It's mathematical concept will be explored and some examples of applying this filter to images will we illustrated.

In this article, State-of-the-art deep learning models are evaluated and their performances in X-ray image classification is reported.

This paper introduces Euclidean distance filter as a new method for image filtering. It discuses conceptual aspects of the filter and implements examples to illustrate it's application.

This is a paper regarding application of deep neural network in prediction of Forex market. It utilized advanced deep learning techniques and software package in order ti evaluate capability of deep neural network in market behavior prediction.