X-ray DR Images Dataset of Camellia Seeds and Chestnuts
Description
The dataset comprises two sub-datasets: one includes preprocessed X-ray DR images of 204 camellia seeds and 237 chestnuts, while the other consists of X-ray DR images acquired through magnification methods, totaling 2040 images of camellia seeds and 2370 images of chestnuts. The camellia seed samples encompass healthy, withered, cavity, and insect-damaged specimens. Similarly, the chestnut samples include healthy, double-kernelled, mold-infested and insect-damaged ones. The X-ray DR images were randomly divided into training, validation, and prediction sets at a ratio of approximately 70:15:15. The training and validation sets were utilized for hyperparameter tuning and classifier training, whereas the independent prediction set samples were employed for external testing of the classifier.