Product Based Classification of Bulk Food Grains using Bag of Visual Words and Deep Features

Abstract

The goal of this research is to compare between the performance of the traditional machine learning classification algorithm using Bag of Visual Words (BoVW) method and off-the-shelf deep features extracted by VGG-19, and Inception-V3 models and trained SVMs using the extracted features. By comparing the AUC, sensitivity, and specificity of SVM with VGG19 and Inception-V3, we can conclude that off-the-shelf deep features has an important impact on food grains image