A High-Resolution Macro-Photography Dataset for Fungal Spoilage Classification on Bread

This dataset contains 314 high-resolution macro-photography images documenting the progression of fungal spoilage on commercial white bread, intended to support the development of automated food safety monitoring systems using computer vision and deep learning. Images were collected by placing bread slices in sealed polyethylene bags with controlled moisture and storing them at room temperature (20–25°C) to simulate realistic household and retail spoilage conditions. All images were captured using an iPhone 17 Pro Max in native Macro Mode at an average resolution of 4016 × 3016 pixels under natural ambient lighting. The dataset is categorized into four morphological classes: Type-Pinhead (132 images, consistent with Rhizopus spp.), Type-Green (92 images, consistent with Penicillium spp.), Type-White (42 images, early-stage mycelium before sporulation), and Mixed (48 images, overlapping multi-species phenotypes). A companion CSV metadata file is provided alongside the images. Data was collected across three locations in Sirajganj, Bangladesh, and annotations were reviewed by personnel with expertise in fungal morphology and food microbiology.