Physics Laboratory Equipment Image Dataset Covering 27 Instrument Categories

<p dir="ltr">This dataset presents <b>3,590 real-world images</b> of physics laboratory apparatus captured at United International University, Dhaka, using four different camera devices. It covers <b>27 equipment classes with 7,024 high-quality annotations</b>, labeled using both bounding-box and polygon formats for geometric precision. Designed for object detection research, the dataset supports training and benchmarking of deep learning models such as <b>YOLOv8 and YOLOv11</b>, which achieved <b>mAP@50 scores of 84.4% and 84.5%</b>, respectively. Beyond research, it enables applications in laboratory safety, automated inventory management, and educational feedback systems, serving as the first open dataset dedicated to physics laboratory equipment detection.</p>