An Optimized Intelligent Deep Network for Herbal Leaf Classification

Abstract

In recent times, a variety of industries have made extensive use of image processing techniques for tasks including segmentation and classification. However, the traditional image processing and ensemble learning approaches face challenges in feature selection and classification. To overcome the demerits of the conventional image processing and boosting algorithm, a novel hybrid Chimp-based Boltzmann Prediction Network (CbBPN) was developed in this article. The presented work was designed and verified in MATLAB software with the herbal leaf dataset. In the model development, the pre-processing and feature extraction module is responsible for extracting valuable features that are pertinent to the classification process. Furthermore, the chimp fitness function increases the classification rate by removing unwanted elements during the classification stage. Additionally, the developed model uses the matching operation to specify the types of the leaf. Furthermore, a case study was created to explain the ways the suggested approach operates. Moreover, a comparison of the projected findings with the existing categorization approaches validates the effectiveness of the constructed model. The comparative analysis shows that the new methods outperformed previously available ones in terms of output.