**Background:** Custard apple (Annona squamosa), also known as sugar apple or sweetsop, is a subtropical fruit valued for its nutritional and medicinal properties, including antioxidants like flavonoids, vitamin C, and lutein, which may protect against chronic diseases and age-related macular degeneration. The fruit is economically significant for the food and pulp industry, but diseases can severely impact yield and quality. Accurate and early disease detection is crucial for effective management. While prior research has explored image-based disease detection in various fruit crops, a dedicated, open-access dataset for custard apple diseases was lacking. This study aimed to fill that gap by creating a comprehensive, high-resolution image dataset of custard apple fruit and leaf diseases to facilitate the development of machine learning-based diagnostic tools.
**Methods:** The dataset was constructed through field surveys conducted in the Nimgaon-Bhogi region, Taluka-Shirur, Pune district, Maharashtra, India (latitude 18.817435, longitude 74.256013) during August to September. High-resolution images were captured using the rear camera of a Samsung Galaxy F23 5G mobile phone (50-megapixel, f/1.8 lens, Sony IMX582 sensor). Images were taken in daylight under natural conditions, as well as after separating fruits and leaves from the plant, to capture diverse disease manifestations. The collected images were reviewed, preprocessed (resized to 768 × 1024 pixels, saved in JPG format at 72 dpi), and classified using IrfanView 64-bit version 4.62 for batch conversion and labeling. Classification parameters included image quality, EXIF data, XMP data, and IPTC data. The final dataset comprises 8226 images organized into six disease categories: Anthracnose (1075 images), Black Canker (1780 images), Diplodia Rot (1645 images), Leaf Spot on fruit (867 images), Leaf Spot on leaf (1255 images), and Mealy Bug (1604 images). The dataset is publicly available on Mendeley Data (DOI: 10.17632/jtgh2885yf.2).
**Key Results:** The dataset provides a diverse and comprehensive collection of custard apple disease images, with a total of 8226 high-resolution images across six categories. The largest category is Black Canker (1780 images), followed by Diplodia Rot (1645 images) and Mealy Bug (1604 images). The dataset includes both fruit diseases (five categories) and one leaf disease category (Leaf Spot on leaf). The images are standardized in resolution (768 × 1024 pixels) and format (JPEG), facilitating their use in machine learning workflows. The dataset is the first open-access resource specifically for custard apple diseases, enabling researchers to train and test algorithms for automated detection and classification.
**Clinical Implications:** Although this is a dataset paper rather than a clinical study, its implications for agriculture and public health are significant. By enabling accurate and automated disease identification, the dataset can help farmers and agronomists implement timely disease management strategies, potentially reducing crop losses and improving yield. This supports sustainable custard apple production, which is important for food security and economic stability in regions where the fruit is cultivated. Additionally, the nutritional benefits of custard apple (e.g., antioxidants, fiber, lutein) underscore the importance of maintaining fruit quality, as diseased fruits may have compromised nutritional value. The dataset's open-access nature promotes collaboration and accelerates research in plant pathology, precision agriculture, and machine learning, ultimately contributing to better crop health and human nutrition.