ABSTRACT

To withstand demand of rice for a massive population worldwide, the detection and classification of rice leaf diseases is significantly important now a days. Usually, rice leaf suffers from numerous bacteriological, virusrelated, or fungous diseases and due to these diseases rice production is gradually decreases. The advancement of convolutional neural network shows the way for detection of rice diseases using local features with the expectation of high returns. In this article, the author propose a CNN-based model to detect and classifying the three different rice leaf diseases using fixed point local features. The achieved results exhibit the efficiency and advantage of our approach in contrast to the state-of-the-art rice leaf disease detection and classification methods.

Keywords: Diseases, Feature Extraction, Local Features, Normalization, Neural Network, Pooling