An Annotated Water-Filled, and Dry Potholes Dataset for Deep Learning Applications

Published: 2 March 2023| Version 1 | DOI: 10.17632/tp95cdvgm8.1
Contributors:
,
,

Description

Dataset Content: - IMG -> Dataset Images. - XML -> Pascal VOC XML Format Annotations. - TXT -> YOLO TXT Format Annotations. - Potholes.zip -> Dataset images and annotation files in compressed format. Video: - Dashcam-Front.mp4 -> A short outdoor front car dashcam video showing the coordinates and car speed to assess the frame rate and accuracy of the detection (faces and plate numbers blurred for privacy). - Dashcam-Rear.mp4 -> A short outdoor rear car dashcam video showing the coordinates and car speed to assess the frame rate and accuracy of the detection (faces and plate numbers blurred for privacy). Note: train.txt and valid.txt files are to be generated based on the location of the dataset files within the training environment.

Files

Institutions

University of Kent

Categories

Image Processing, Object Detection, Object Recognition, Image Classification, Convolutional Neural Network, YOLOv5, Semantic Neural Network

Licence