Answers are grounded in this paper and cite their sources.
Rice is a vital staple crop for millions of people in the world and increased productivity is significant for food security. Nevertheless, rice plants are broadly prone to several leaf diseases that severely impact the crop if it is not detected early. Conventional disease identification techniques mainly depend on manual inspection which is labor-intensive, and time-consuming, particularly in large-scale farming. These issues are occurred due to the lack of an accurate and automated systems. To tackle these problems, an effective rice plant leaf disease classification framework is proposed named Harmonic Euclidean Smooth Tamed Loss (HaEST-Loss) for Pyramid Dilated Kronecker Network (PDKN). The input image is pre-processed by using contra-harmonic mean filter. Then, LadderNet is used for segmenting the leaves. After that, feature extraction is performed from segmented plant leaf image. Thereafter, leaf disease classification is conducted by PDKN with HaEST-Loss. Here, HaEST-Loss is designed by the integration of using Hormonic analysis, Tamed Cross Entropy (TCE) Loss, Euclidean loss and smooth surrogate loss for PDKN. The efficiency of HaEST-Loss is analyzed by considering performance metrics, like Accuracy, True Positive Rate (TPR) and True Negative Rate (TNR), which acquired the values of 91.646%, 90.728% and 92.610%.
No takes yet. Share an insight, caveat, or question.
V et al. (2026) studied this question.