Spiny Gourd Tri-Class Quality Dataset
Description
The "Spiny Gourd Tri-Class Quality Dataset" is a specialized, curated subset extracted and refined from the open-source "Nature's Best vs. Bruised: A Veggie Evaluation" dataset under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This dataset focuses specifically on Spiny Gourd (Momordica dioica), a widely consumed vegetable, to facilitate automated quality assessment and edibility evaluation using Computer Vision (CV) and Deep Learning (DL) models. Key Details of the Dataset: 1. Target Crop: Spiny Gourd (Scientific Name: Momordica dioica) 2. Classification Categories: The dataset is structured into three distinct physical quality classes: - Fresh: High-quality, unblemished vegetables suitable for consumption. - Damaged: Vegetables showing minor structural, mechanical, or surface defects. - Severely Damaged: Highly degraded or rotten specimens unsuitable for commercial sale/consumption. 3. Environmental Conditions: Images feature authentic natural progression captured under varied camera angles, non-uniform illumination conditions, and consistent background setups. Purpose & Academic Usage: This dataset is designed to address the lack of specialized single-crop quality benchmarks in agricultural computer vision. Researchers can utilize this dataset to train, validate, and benchmark deep learning architectures (e.g., CNNs, Vision Transformers) for real-time agricultural sorting, post-harvest quality inspection, and automated grading systems. Attribution & License: Original Dataset: "Nature's Best vs. Bruised: A Veggie Evaluation" License: Creative Commons Attribution 4.0 International (CC BY 4.0) Modification Note: Filtered specifically for Spiny Gourd samples, re-annotated, and structured into a three-class quality taxonomy.