Logo Based Amphetamines Classification using SURF and Bag-of-features model
Tharaphon Nitijiramon, Nagul Cooharojananone, Somjet Saiseng · 2020
In this paper, we propose a framework for classifying the top view image of amphetamines based on their logo using SURF and Bag-of-features model. During our experiment, we found that the unsmooth surface of amphetamines and low contrast are the main factors of low accuracy for classification. Therefore, we propose a process to enhance the main feature and reduce noise on the surface using adaptive filter, Contrast-Limited Adaptive Histogram Equalization (CLAHE), active contour and image morphology. The result from our proposed preprocess algorithm shows that the clarity of the logo on amphetamines is improved and the noise is reduced. We also then apply SURF to extract features and classify using Bag-of-features model. This experimental result shows that our proposed preprocess for each step can improve the accuracy up and the accuracy of our method up to 97 percent.