Plant Disease Detection in Imbalanced Datasets Using Efficient Convolutional Neural Networks With Stepwise Transfer Learning

Abstract

Convolutional neural networks have demonstrated state-of-the-art performance in image classification and various other computer vision tasks. Plant disease detection is an important area of deep learning which has been addressed by many recent methods. However, there is a dire need to optimize these solutions for resource-constrained portable devices such as smartphones. This is a challenging problem because deep learning models are resource extensive in nature. This paper proposes an efficient method to systematically classify plant disease symptoms using convolutional neural networks. These networks are memory efficient and when coupled with the proposed training configuration it enables rapid development of industrial applications by reducing the training times. Another critical problem arises with the improper distribution of samples among classes known as the class imbalance problem, which is addressed by employing a simple statistical methodology. Transfer learning is a known technique for training small datasets which transfers pre-trained weights learned on a large dataset. However, during transfer learning, negative transfer learning is a common problem. Therefore, a stepwise transfer learning approach is proposed which can help in fast convergence, while reducing overfitting and preventing negative transfer learning during knowledge transfer across domains. The system is trained and evaluated on two plant disease datasets i.e., PlantVillage (a publicly available dataset) and pepper disease dataset provided by the National Institute of Horticultural and Herbal Science, Republic of Korea. The pepper dataset is particularly challenging as it contains images from different parts of the plant such as the leaf, pulp, and stem. The proposed system has dominated the previous works on the PlantVillage dataset and achieved 99% and 99.69% accuracy on the Pepper dataset and PlantVillage datasets, respectively.

Leaf Diseases Identification and Classification of Self-Collected Dataset on Groundnut Crop using Progressive Convolutional Neural Network (PGCNN)

“…Most of the dataset used for model building are taken from publicly available dataset. The study [21] also uses both the self-collected dataset and PlantVillage dataset for pepper leaf disease detection. Groundnut leaves are collected from the various fields located near the village of Pudukkottai district and also from various websites through google search engine.…”

Smart Irrigation for Crop Management Using IoT

“…Tests show that the model can successfully diagnose a set of data and the accuracy of overall recognition which is why it can be used effectively in diagnosing and diagnosing plant and insect diseases. Paper [5] proposes an effective way to systematically classify the symptoms of plant diseases using neural convolution networks. This program is trained and tested on two plant disease databases namely plant data and pepper data sets.…”

LSCEC: Design of a High-efficiency Leaf-image Segmentation & Classification Model via Ensemble Compute Processes

“…Models that use high density CNNs [14], Stepwise Transfer Learning [15], Dual Attention with Topology Fusion Generative Adversarial Networks [16], Data Augmentation with Generative Adversarial Networks [17], Deep CNN [18], image compression-based CNN [19], and Restructured Deep Residual Dense Network [20] are also proposed by researchers. These models showcase high-density feature extraction capabilities, but have higher complexity, which reduce their feature representation capabilities.…”