ECTI TRANSACTIONS ON COMPUTER INFORMATION TECHNOLOGY

Volume 14, No. 01, Month MAY, Year 2020, Pages 1 - 10

A comparative study of rice variety classication based on deep learning and hand-crafted features

Vinh Truong Hoang, Duc Phan Van Hoai, Thongchai Surinwarangkoon, Huu-Thanh Duong, Kittikhun Meethongjan

Abstract Download PDF

Rice is vital to people all around the world. The demand for an efficient method in rice seed variety classification is one of the most essential tasks for quality inspection. Currently, this task is done by technicians based on experience by investigating the similarity of colour, shape and texture of rice. Therefore, we propose to find an appropriate process to develop an automation system for rice recognition. In this paper, several hand-crafted descriptors and Convolutional Neural Networks (CNN) methods are evaluated and compared. The experiment is simulated on the VNRICE dataset on which our method shows a significant result. The highest accuracy obtained is 99.04% by using DenNet21 framework.

Keywords

Color Texture, Texture Classication, Deep Learning, Hand-crafted Feature, LBP, HOG, SIFT, GIST, Rice Seed Classication

---