Auto-Tune Dialog

The Auto-Tune dialog is used in conjunction with the OCRMax function to configure the function's Segmentation parameters and train a font database, through the construction of Auto-Tune Records to create an Auto-Tune Model. An Auto-Tune Record consists of an image, the training string and the user-verified location of the characters within the image. The Auto-Tune Model consists of one or more Auto-Tune Records, plus the settings that best segments all of the images within the Auto-Tune Records.

Note:
  • When training a font database and using the Auto-Tune dialog to segment the characters, Cognex recommends beginning the process with images that best represent how the characters should look (i.e. use your best images first), working from best to worst. Also include images that represent all of the variations that will be expected to be encountered during run-time. This will allow the OCRMax function's Auto-Tune algorithm to tune more successfully.
  • The Auto-Tune algorithm will not work on text that exhibits an angle larger than +/- 5 degrees within the ROI. In this scenario, either ensure that the region of interest (ROI) is fixtured so that it is aligned with the text within the angle range, or set the Angle Range parameter in the OCRMax Segmentation tab to a value greater than 5, at which point the Auto-Tune algorithm will utilize the value in its calculations.
  • The Auto-Tune algorithm will not work when characters are overlapped. Perform manual segmentation to segment and classify overlapped characters.

Overview

With the Auto-Tune dialog running, each image is examined to verify that the characters are being correctly segmented and classified. If the characters are not being correctly segmented, the OCRMax function's Auto-Tune algorithm will calculate the optimal Segmentation parameter settings that segment the current image (the Auto-Tune Record), as well as the previously trained images (the Auto-Tune Model). As more images are trained, the OCRMax function's Auto-Tune algorithm will become more reliable and accurate. Once satisfactory results are achieved, the Auto-Tune dialog is closed, the new Segmentation parameters are applied and the font database is updated with the newly trained characters.

The Auto-Tune dialog provides two primary advantages over manually tuning the Segmentation parameters:

  1. Images only need to be cycled through once, instead of twice (once to tune the Segmentation parameters, and the second to train).
  2. Read accuracy should improve because the characters are trained automatically with the Segmentation parameters obtained during the tuning process.