Object classification using a single-pixel camera and neural networks

Single pixel imaging is a promising method of imaging without using multi-pixel matrices. Unlike traditional methods, the image here is not directly registered, but computed. Recently, machine learning techniques have started to be used to solve this problem. In this paper, we show the potential application of convolutional neural networks in single-pixel imaging to classify objects from a substantially incomplete set of measurements. We find the dependence of classification accuracy on various object sampling parameters. The proposed methods can be used in real devices as efficient software.