dumuzeyn/ROdataset

The dataset consists of 1109 images showing traffic lights, cars, buses, trucks, and road signs. The dataset is marked up in CVAT.AI . There are 4 classes in the dataset: Car, Person, RoadSigns, Bike.

1. The CAR — class refers to all types of transport on the roads except trains.
2. The Person — class refers to people who are near the road, unfortunately there are minor problems with this class. By training the yolov8n neural network and other types of YOLO, I deduced the problem that I don't always see people not in black clothes. This problem is related to the fact that the footage was taken in winter in Russia, which is why all the people in the pictures are wearing black jackets.
3. The Bike — class refers to motorcycles and bicycles, but the dataset only has this class in 5 pictures, so we can assume that the dataset does not have this class.
4. The RoadSigns — class includes all road signs, parking signs, and traffic lights.

To install the dataset, use this command:

from datasets import load_dataset
ds = load_dataset("dumuzeyn/ROdataset")