Number: 046, Date: 2024-11-25

2024.11.19 ~ 2024.11.24

๐Ÿ‘๏ธย Real-Time Text Recognition

If you want to recognize text in real-time, iOS provides frameworks like Vision and VisionKit.

Depending on the iOS version, there are several implementation methods:

iOS 16 and above

  1. Using VNDocumentCameraViewController โ†’ Custom UI is difficult to implement due to its limitations.
  2. Using DataScannerViewController โ†’ Offers various options with LiveText (e.g., phone numbers, emails, URLs, etc.).

iOS 16 and below

  1. Using AVFoundation + Vision โ†’ Combines image capture and video processing from AVFoundation with text recognition using Vision.

After implementing these methods, I found that LiveText is advantageous for its ability to highlight recognized text and for its simplicity. On the other hand, the AVFoundation + Vision approach is faster and compatible with lower OS versions.

๐Ÿ™‹๐Ÿปโ€โ™‚๏ธ Other Notes

  1. When using Local Packages with Tuist, running Fastlaneโ€™s increment_build_number lane requires resetting the package cache every time. I have yet to find a solution for this issue.
  2. git cherry-pick is useful for applying specific commits, but it has a high risk of conflicts. Thus, using it on the main branch requires extra caution.
  3. I wrote a blog post about the open-source project SocketClusterNative.