Scopus Indexed Publications

Paper Details


Title
Machine Vision-Based Image Recognition: A Curated Image Dataset

Author
Mohammad Monirul Islam,

Email

Abstract

This study presents a curated paddy pest image dataset for deep learning-based computer vision classification. The dataset includes 22,108 images of six major paddy pest classes. In total, 2,557 original images were captured in real paddy field environments using smartphone cameras under different lighting conditions, backgrounds and pest orientations. To increase the diversity of the dataset and the generalization of the model, 19551 augmented images were created using rotation, scaling, cropping, brightness adjustment and horizontal flipping. Each image was carefully labeled and verified with the help of experts to ensure the reliability of the annotations. The dataset was also evaluated in terms of important quality parameters like relevance, diversity, annotation accuracy, size of the dataset, quality of preprocessing and usability of the model. For automated pest classification, the processed dataset was used to train and test a Custom Convolutional Neural Network to validate its effectiveness. The experimental results showed that the dataset is suitable for smart pest monitoring, and the classification accuracy was high for all six classes of pests. The proposed dataset can be used for precision agriculture, early pest detection, reduced pesticide usage and sustainable rice production.


Keywords

Journal or Conference Name
Proceedings of the 3rd International Conference on Innovations in Cybersecurity and Data Science, ICICDS 2026

Publication Year
2026

Indexing
scopus