CODEWITHSUNDEEP
iosswiftapple-visiontext-recognition
Alcibiades
Alcibiades

Reputation: 435

VNRecognizedText returns only 3 confidence values

I'm using VNRecognizeTextRequest from iOS's Vision framework to detect text in images.

The confidence that I'm getting for all the detected text is always one of these 3 values: 0.3, 0.5, 1 I've verified this on thousands of images, using both .fast and .accurate options.

The documentation states that confidence should be normalized between 0 and 1, and not a set of 3 values:

The confidence level is a normalized value between 0.0 and 1.0, where 1.0 represents the highest confidence.

Is anyone else experiencing this? Is it expected? If you're getting more diverse confidence values, please share your code snippet and the values you're getting.

Thanks!

Upvotes: 2

Views: 97

Answers (0)

Related Questions