DoR

Scopus Indexed Publications

Title: Malabar Nightshade Disease Detection and Recognition Using Transfer Learning

Abstract: As an agricultural country it is of utmost important for us to ensure the maximum production from this sector as the local demand is increasing day by day. Malabar nightshade leaves which are produced locally are often prone to different kinds of scabs which seriously affects the overall production of the same. In this a paper a model has been proposed to distinguish healthy Malabar leaves from the scabs. MobileNet and Xception are used to train the model for the purpose of predicting the disease related with Malabar leaves. So far there is very few research in this area. As a result, there is ample scope to explore and improve the detection and classification process. The mentioned pre-trained models are more effective considering the traditional one. We have achieved an accuracy of 96.77% using our proposed model which is very effective. While comparing, it is found that both algorithm's confusion matrices are also the same. So we have used the loss value instead after training the model which proves the strength of MobileNet compared with other available models.

Keywords: Transfer Learning , Mobilenet , Xception , Mal-abar Leaves , Scab on Leaves , Leaves Disease Detection , Disease Classification

Journal or Conference Name: 2021 IEEE 6th International Conference on Computing, Communication and Automation (ICCCA)