Fast Image Classification Algorithms Based on Sequential Analysis

This work is devoted to fast image recognition techniques based on statistical sequential analysis. We examine the possibility sequentially process principal components and organize a convolutional neural network with early exits. pay special attention learning multi-task lightweight model in order predict several facial attributes (age, gender, ethnicity) preliminary training face classification task. specifically note that entire above-mentioned should be fine-tuned deal emotion problem. Experimental evaluation datasets demonstrates proposed approach has high accuracy very low running time space complexity compared known state-of-the-art methods.