DPCR_Image_data_sheet_for_yolo

This dataset is specifically designed for target detection and analysis in digital polymerase chain reaction (dPCR) workflows. It contains high-resolution droplet images acquired from seven fluorescence channels, accompanied by standardized object detection annotations. The dataset consists of two main components: 1. Refined Droplet Dataset (YOLO Format) This component contains 3,584 images focusing on droplet regions extracted from seven fluorescence channels. All images have been fully annotated for object detection tasks. To facilitate seamless model training, the dataset has been preprocessed and organized according to the standard YOLO directory structure, with each split containing corresponding images/ and labels/ subfolders. Training set: 2,688 images (75%) Validation set: 448 images (12.5%) Test set: 448 images (12.5%) 2. Raw Full-Size Images and Inference Results To support large-scale image processing and deployment performance evaluation, the dataset also includes: Raw dPCR Images: Original ultra-high-resolution digital PCR images (2448 × 10000 pixels). Inference Outputs: Complete object detection and classification results mapped directly onto the corresponding full-size raw images, serving as practical benchmarks for deployment-level performance evaluation.