Testing strength of the state-of-art image classification methods for hand drawn sketches
Ochilbek Rakhmanov · 2019
Classification of hand drawn sketches (images) reached a classification accuracy of % 77 with the latest state-of-the-art method, called Sketch-a-Net, in 2017. Most of the developed methods use image feature extractor techniques like HOG, BOVW, or CNN. In this paper, we tested the classification accuracy of hand drawn sketches with SVM and ANN, without using image feature extraction algorithms and compared the results with the findings of a number of important state-of-art researches. Our findings show that existing methods are reasonable to accept, even though the results of our experiments also produced some valuable results. We propose that our findings can serve as kind of 'minimal milestone' on future prediction experiments.