Sample Segmented Wheat Grain Dataset

Summary (short field):A publicly available sample dataset of segmented wheat grain images from the CGrain system, containing 10 representative images per class for reproducibility and validation in CNN-based classification. Full Description (main field): This dataset contains sample segmented wheat grain images used in the study “Investigating Performance and Key Factors for Real-World Deployment of Grain Image Classification Using Convolutional Neural Networks.” The images were acquired using the CGrain mirror-based imaging system and processed with its proprietary pipeline to generate 256 × 256 pixel segmented images. The dataset includes 10 representative samples from each of the eight wheat quality classes (Sound, Broken, Black Germ, Fusarium, Moldy, Sprouted, Spotted, and Insect), collected from Instrument 864. These samples are provided in accordance with peer reviewer requests to enhance transparency, reproducibility, and visual interpretability of the experimental setup. Due to industrial confidentiality, only a limited subset is publicly shared, and detailed preprocessing algorithms are not disclosed. Please cite this dataset as:Kumari, R. (2025). Sample Segmented Wheat Grain Dataset [Data set]. Zenodo. https://doi.org/10.5281/zenodo.17397123