# A comprehensive review on Plant Leaf Disease detection using Deep learning

**Authors:** Sumaya Mustofa, Md Mehedi Hasan Munna, Yousuf Rayhan Emon, Golam Rabbany, Md Taimur Ahad  
**arXiv:** [2308.14087](https://arxiv.org/abs/2308.14087)  
**Date:** 2023-08-27  
**Summary page:** https://www.summarizepaper.com/en/paper/2308.14087v1/

## Key points

- Leaf disease is a common and potentially fatal issue for plants
- Early diagnosis and detection are crucial for improving prognosis
- Several automated systems utilizing different plant pathology imaging techniques have been developed to predict leaf diseases
- This paper presents a comprehensive review of deep learning models used for diagnosing various plant leaf diseases
- Deep learning models discussed in the review include Vision Transformer (ViT), Deep Convolutional Neural Network (DCNN), Convolutional Neural Network (CNN), Residual Skip Network-based Super-Resolution for Leaf Disease Detection (RSNSR-LDD), Disease Detection Network (DDN), and You Only Look Once (YOLO)
- These models have been applied to publicly available datasets to compare their performance using metrics such as accuracy, precision, recall, etc.
- Borhani et al. (2022) developed a lightweight deep learning-based method using ViT combined with conventional CNN techniques and attention blocks for automated plant disease categorization
- Bandi et al. (2023) proposed a model based on YOLOv5 for categorizing and detecting plant leaf disease stages by removing the background of the diseased leaf using image processing techniques
- The review highlights advancements in utilizing deep learning models for leaf disease detection
- Emphasizes the need for accurate diagnosis methods to mitigate the impact of these diseases on plants

## Summary

Leaf disease is a common and potentially fatal issue for plants, making early diagnosis and detection crucial for improving prognosis. In recent years, several automated systems utilizing different plant pathology imaging techniques have been developed to predict leaf diseases. This paper presents a comprehensive review of the literature on deep learning models used for diagnosing various plant leaf diseases. The review discusses the advantages and limitations of different deep learning models employed in leaf disease detection, including Vision Transformer (ViT), Deep Convolutional Neural Network (DCNN), Convolutional Neural Network (CNN), Residual Skip Network-based Super-Resolution for Leaf Disease Detection (RSNSR-LDD), Disease Detection Network (DDN), and You Only Look Once (YOLO). These models have been applied to publicly available datasets in order to compare their performance using metrics such as accuracy, precision, recall, etc. The study identified 540 candidate articles from Google Scholar and Scopus databases. Among these studies, Borhani et al. (2022) developed a lightweight deep learning-based method for automated plant disease categorization using a vision transformer called ViT. The authors combined conventional CNN techniques with attention blocks to speed up prediction. Another study by Bandi et al. (2023) proposed a model for categorizing and detecting plant leaf disease stages based on the severity of infection. They utilized You Only Look Once version 5 (YOLOv5) deep learning model to detect plant leaf diseases. The background of the diseased leaf was then removed using image processing techniques. Overall, this review highlights the advancements made in utilizing deep learning models for leaf disease detection and emphasizes the need for accurate diagnosis methods to mitigate the impact of these diseases on plants.