Object Detection Benchmark Dataset for Indoor Power Sockets and Light Switches

Read full text Summarise with VIRAT
Authors: Lukas Arzoumanidis, Falko Behre, Malte Marsmann, Weilian Li, Youness Dehbi
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

We merged different datasets from Roboflow and converted them to the YOLO annotation format, enabling the use of the Ultralytics ecosystem of pretrained YOLO models and facilitating rapid deployment of custom-trained models on mobile devices. The combined dataset form Roboflow comprises 3,459 images distributed across the three classes. The majority of samples (2,593) correspond to power sockets, while the smallest subset (151) belongs to the power strip class. All images in the prepared dataset exhibit varying resolutions, lighting conditions, and orientations. The prepared dataset was partitioned into 80% for training, 10% for testing, and 10% for validation. Datasets used:https://universe.roboflow.com/test-gxlza/power-strip-cosfvhttps://universe.roboflow.com/test-gvutb/socket_rockethttps://universe.roboflow.com/yolov5-power-socket-detection/power-socket-detection-rofcvhttps://universe.roboflow.com/timengelbracht/spotlight-light-switch-dataset

Fields of Study & Subjects

Advanced Neural Network Applications Smart Agriculture and AI Industrial Vision Systems and Defect Detection

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
1 April 2026
Language
EN
Document type
Dataset