'Does Vision framework based barcode scanner work best with a particular video size/ pixel format?
When building a Vision framework based barcode scanner, A colleague was asking why I set AVCaptureSession sessionPreset to .hd1280x720 and output pixel format type to kCVPixelFormatType_32BGRA rather than using a higher resolution/bit depth - why not 4K 64bit video?
In truth, I copied these values from another implementation My assumption is that the Vision algorithm or underlying CoreML model (if its using one) was trained or optimised to work with images at this resolution/bit depth. But I haven't found any documentation to support this.
Does anyone know if my assumption is correct, is the Vision framework barcode scanner designed to be used with a specific size/format pixel buffer?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|