I have a set of 52 images of playing cards, each card goes through some image processing to distort the cards in various ways, resulting in around 1200 images. Each image file contains only the picture of the card and nothing else.

Is there a simple way to label this without having to use something like labelme and annotating the image.

Basically what I want is to attach a label to the file name, rather than annotating it by hand. So that for example, all the Ace image files are labelled Ace without me doing it manually.

I am not opposed to switching my ML model to something other than Yolov5

CodePudding user response：

The only thing that you can do to save time is to label a subset manually train the model and then use this model to run inference on unlabelled datasets and get boxes and save them as annotation files for each image.