PalmRachis-BiLSTM-Attn: An Anatomically Guided Explainable Deep Learning Framework for Spatial–Temporal Progression Modelling of Date-Palm Leaf Diseases

Project Code :TCMAPY2374

Objective

The main objective of this project is to develop a robust and explainable deep learning framework for the classification and detection of date-palm leaf diseases. This will be achieved by utilizing advanced algorithms such as PalmRachis-GAT-BiLSTM and PalmRachis-BiGRU-Attn, which analyze both spatial and temporal data from date-palm leaf images. The project aims to enhance the accuracy of disease classification, focusing on diseases such as Potassium Deficiency, Fusarium Wilt, and other common leaf diseases. A key focus is on creating an explainable model that provides transparent results, making it easier for agricultural professionals to interpret the system's decisions. The project will also develop a modular and user-friendly system, including features like user registration, disease classification, and session management, to ensure accessibility for farmers and agricultural experts. Additionally, the model will be tested and validated on large datasets to guarantee its reliability across various conditions and disease types. Ultimately, the project aims to contribute to improved agricultural practices by offering an automated solution for early disease detection, which can lead to better crop management, increased productivity, and more informed decision-making.

Abstract

Date-palm trees are a vital agricultural asset in many regions, providing significant economic value. However, diseases affecting their leaves pose a major threat to their growth and productivity. Detecting and diagnosing these diseases at an early stage can substantially improve crop management and yield. This research presents an anatomically guided explainable deep learning framework designed for spatial-temporal progression modeling of date-palm leaf diseases. The model utilizes PalmRachis-GAT-BiLSTM and PalmRachis-BiGRU-Attn algorithms to classify and detect leaf diseases by analyzing both spatial and temporal patterns in the images of date-palm leaves. By leveraging graph attention networks (GAT) and BiLSTM, the system can handle complex spatial-temporal features while providing explainability for better decision-making. The system’s modular approach includes user-friendly interfaces for registration, login, disease classification, and logout functionalities, developed using HTML, CSS, JavaScript, and Flask. The model's high accuracy ensures effective disease detection, with potential for application in precision agriculture.

Keywords: Date-palm diseases, Deep learning, BiLSTM, GAT, Disease classification, Spatial-temporal, PalmRachis, BiGRU, Machine learning, Agriculture.

NOTE: Without the concern of our team, please don't submit to the college. This Abstract varies based on student requirements.