Cassava Syndrome Scan a Pioneering Deep Learning System for Accurate Cassava Leaf Disease Classification

Cassava serves as a crucial source of carbohydrates for a vast population globally, playing a significant role in food security. The crop, however, faces threats from a variety of leaf diseases, leading to substantial reduction in yield. Prompt and precise identification of these diseases is essential for implementing effective countermeasures and maintaining adequate food supply. Recently, deep learning methodologies have demonstrated remarkable capabilities in image-based recognition and classification, including the diagnosis of plant diseases. The objective of this research is to construct a dependable cassava leaf disease identification system utilizing deep learning approaches, with an emphasis on achieving exceptional accuracy and dependability. We assess the performance of multiple cutting-edge deep learning architectures, such as ResNet50, DenseNet121, EfficientNet-B0, NasNetLarge, and EfficientNet-V2, in the context of cassava leaf disease classification. The dataset is comprised of nearly 15,000 test images, accompanied by image_id and label data in train.csv and sample_submission.csv files. Data preprocessing, including resizing, normalization, and augmentation, is conducted before partitioning the data into training, validation, and test sets. Our findings indicate that the EfficientNet-V2 model surpasses its counterparts in performance, attaining the highest validation accuracy of 0.7480 and a comparatively low validation loss of 0.7488. We further explore several strategies to enhance model performance, such as optimizing hyperparameters, employing data augmentation, utilizing early stopping, implementing learning rate scheduling, applying transfer learning, and adopting ensemble techniques. The resulting classifier offers potential benefits to farmers and agricultural specialists in making informed decisions regarding cassava leaf disease management, ultimately leading to improved crop yields and food security.