Description
Roadsigns images, sampled around Manchester. Each image is cropped close to the road sign and is 256x256 pixels. There are 31 different road signs, each in its own directory and a total of 55 000 images. 213 videos were sampled containing one or more roadsigns each. Different distances, angles, rotation, exposure, and zoom levels were captured for each video to get variations in the data. Zoom helped due to using a phone with 3 cameras making sure each video used all of its camera sensors to get some extra variation. The images were then sampled using a script at different sampling frequencies according to the total sampled data for the class. However, the dataset is not even, which is something one must consider if using this dataset. I verified all images, so there should not be any images that are not correct. However, some images are cropped a bit further out and some closer than they should be, but I decided to keep them as they still represent the class, and any human would recognise them.
Citations (0)
No citations found
Mentions (0)
No mentions found
Metrics Over Time
Publication Details
DOI
Publisher
Zenodo
Subfield
Language and Linguistics
Field
Arts and Humanities
Domain
Social Sciences
Confidence Score
84%
Source
Open Alex