AI BriefingCase Study: Improving Barcode Scan Performance in the Olive Young App's Smart Scanner
Key point
Olive Young improved its smart scanner feature by adopting iOS's Vision framework and Android's MLKit to boost barcode recognition rates.
Details
Olive Young carried out technical improvements to address the low barcode recognition rate of its Smart Scanner feature, an issue that occurred during the 'APP-ppun Festival' event. The main causes identified were small barcode sizes, low contrast ratios, and light reflection off packaging.
On iOS, pinch-zoom functionality using UIPinchGestureRecognizer was added to the existing AVCaptureSession approach to improve handling of small barcodes. In addition, to maximize recognition rates, the Vision framework's VNDetectBarcodesRequest was introduced, enabling direct analysis of image pixel buffers.
On Android, the existing ZXing library was removed, and a new structure combining Google's CameraX and MLKit was adopted. This resolved UI customization constraints, and MlKitAnalyzer was used to optimize real-time image analysis performance.