Medicinal plant classification using attention mechanism driven VGG19 and Resnet101

The accurate classification of the medicinal plants is essential in various fields such as pharmaceutical research, traditional medicine, and biodiversity conservation. In this study, we focus on advancing the performance of two popular convolutional neural network architectures, VGG19-BN and Resnet101, for the classification of plant species which have medicinal properties. We propose the integration of attention mechanisms, specifically the Channel-wise Attention Module (CBAM) and the Efficient Channel Attention (ECANet), to improve the feature representation and discrimination capabilities of the models. By enhancing the attention to salient regions, the models become more adept at capturing intricate botanical attributes that contribute to accurate classification. Overall, this study contributes to the advancement of state-of-the-art methods for medicinal plant classification through the integration of attention mechanisms.