Rice Variety Classification in ICETIR 2020

Introduction

Rice is the main source of staple food in the world. Novel methods of rice classification are needed to support food industry, government, and consumer as modern replacements for physical evaluation method employed in many industries. Image recognition offers a cost effective and scalable technology for rice classification. In this research we use deep learning technology by transfer learning method from VGG16 model to classify rice varieties in bulk samples. The best trained model accuracies were 95,56%. Our results show that the transfer learning approach for image recognition of field images offers a fast, affordable, and easily deployable strategy for digital rice classification.

@inproceedings{ICETIR},
  author    = {I. Rosyadi and V.H. Hidarlan and F. Asriani and R. Ediati},
  title     = {Rice Variety Classification},
  booktitle = {ICETIR 2020},
  pages     = {489--503},
  year      = {2020}
}