Herbal Leaf Classification using RCNN, Fast RCNN, Faster RCNN

Herbal leaf classification is an important task in botany and pharmacology with the increasing demand for natural remedies, accurate identification and classification of herbal leaves is crucial for their safe and effective use. In recent years, advances in digital imaginative and system studying techniques has provided promising results in automating this task. The focus of this work is to provide a thorough comprehensive examination on herbal leaf, which is manually collected 4 different classes each class contains 2 different subclasses that is Betel leaf- Mysore Betel leaf and Ambadi Betel leaf, Tulsi- Rama Tusli and Krishna Tulsi, Spinach- Red Spinach and Green Spinach, Hibiscus- Multi Layer Hibiscus and single layer hibiscus. Then, we tried out the existing hierarchical learning models including CNN, RNN, LSTM, and BILSTM to determine better extract and also we used RCNN, Fast RCNN, Faster RCNN object detection algorithm with model relevant features from herbal leaf images, leading to improved accuracy and speed of classification with comparative study. Overall, this paper contributes valuable insights into the current state of research which only considers a limited number of deep learning models such as Convolution neural network, RNN, BILSTM along with Long Short-Term Memory which did not compare a suggested approach to other modern day techniques or models. So, a comparative analysis with other methods like RCNN, Fast RCNN and Faster RCNN are all object detection algorithms based on deep learning that can be used to classify leaf images which are all able to detect objects even when they are partially occluded, making them effective for identifying individual leaves will help to assess the success of the suggested method is to figure out its effectiveness and its limitations on herbal leaf classification.