Crop image

If you want to use the "Crop image" function, please select it when creating a new model.

The image cropping function only supports object detection and semantic segmentation, and does not support OCR. It is suitable for scenes where the images are large but the target objects are small, such as detecting small spots or small cracks in industrial images.

This function not only crops the training images, but also automatically crops the testing images.

Please pay attention to the requirements for cropping the images when uploading.