Image/Video Text Extraction

Extract text and numbers from image and video frames

The component will process each image or video frame sent to it's input to identify and extract text. The component is configurable to output either text or numbers. You can customize the component with regex to identify something specific.

Configuration

The following configuration options are available

Search regex value

You may use regex to search the detections and output specifically what you are looking to identify.

Regex format

We suggest using Regex101.com to build your regex. Once your regex is built navigate to Code generator > Swift 5.2

let pattern = #"^([0-9]{6})$"#

Copy the value in between #" "#

Value pasted into the Cloneable component options ^([0-9]{6})$

Example

Extract exactly 6 digits ^([0-9]{6})$

Only output numbers

The component will attempt to cast the detected text to a number. This will only output the value if it is successful in converting the text to a number. No error will be shown if it's unable to.

Number of verifications

In order to reduce false positive results which may be incorrect you may select the minimum times the identified text must be recognized before outputting the result. The default and suggested value is 5. A higher value will result in a more accurate identification, however will take longer to output the result.

Example use in app

A great use case for easily extracting text is to pair it with a camera view app that is configured to have a text input box. You can feed the video feed into the text extraction component and feed the output into the text box input in the camera app. This will allow for real time text extraction.