Deep learning software Mech-DLK V2.5.0 released, with unsupervised segmentation, OCR, etc

Mech-DLK V2.5.0 officially released, featuring new algorithm modules for new user experience:

  • Unsupervised Segmentation: The model trained with OK images only can realize pixel-level detection of all known and unknown defects, helping clients quickly find the defective workobjects.
  • OCR—Text Detection: Locates texts of specified regions and multiple lines.
  • OCR—Text Recognition: Recognizes the characters present in the image, including numbers, letters, and some special symbols.

Download the latest Mech-DLK here

Usage instructions