Blue Read Tool

The Blue Read tool is used to perform Optical Character Recognition (OCR). It is a pre-trained tool, having been trained on a large database of images with text in numerous fonts and backgrounds. The characters of the pre-trained tool are numerals (0-9), uppercase letters (A-Z, except "O"), hyphen (-), plus sign (+), ampersand (&), colon (:), and forward slash (/).

Similar to the Blue Locate Tool, the Blue Read tool identifies and locates characters as features in the image. However, as noted above, the Blue Read tool is a pre-trained tool, which provides it with a generic baseline for reading performance, without the need for training. Thus when the tool is first configured, it has the ability to almost instantly recognize and read characters. The tool already knows how to read characters; you just need to define where in the image to look for characters.

The Blue Read tool's strength is in its ability to work on difficult projects, i.e. low contrast, low resolution and/or deformed characters. In addition, there are no difficult to set segmentation settings or image filters required. The tool is able to read characters that traditional machine vision tools had difficulty reading; particularly deformed and/or curved characters on noisy backgrounds.

To use the tool, you provide a Training Set and then set the region around the characters to be read. Adjust the Feature Size parameter and label the characters. In most instances, the tool will automatically identify and correctly read the characters, and produce markings of the characters, which you can accept as labels for further training. Once at least one instance of the characters are labeled, train the tool. Then validate the tool on images that were not used during the training phase.