This dataset contains 11,070 images and their corresponding labels manually annotated by me. I have used makesense.ai to annotate these images, and have ordered them in 27 classes corresponding to the clothes. I have applied 10 augmentations to the images.
There will be an images and labels folder inside the branch. Inside each there will be train and val. You may change the number of datapoints (images) in the folders, but make sure to do so for both the images and labels. In accordance to conventions, each image will have a label with the same name in the folder order mentioned. I have also provided the corresponding .yaml file that can be fit into YOLOv5 or YOLOv7 model.
The novelty here is to include "Indian" clothes that are worn by the people, in addition to the already widespread availability of the conventional clothes. I have included Kurtis, Kurtas, Nehru Jackets, Lehengas, etc in the process of annotations.
Here is the exhaustive list of classes used :
0: tshirt
1: shortsleeveshirt
2: longsleeveshirt
3: blazer
4: lehengas
5: jeans
6: trousers
7: shorts
8: capris
9: sarees
10: kurtis
11: kurtas
12: skirts
13: tunic
14: leggings
15: croptop
16: salwaar
17: vests
18: hoodies
19: joggers
20: nehrujackets
21: jumpsuit
22: playsuit
23: dresses
24: shrugs
25: jacket
26: top