Release 1.7.1

⚠️ Breaking Changes

• Orientation is handled by frame orientation. See Orientation docs for more details.
• If you want to use autoScale prop you should define windowWidth and windowHeight in faceDetectionOptions.
  With this you can define if you want to remove status bar size from screen size, etc...

✨ Features

• add windowWidth and windowHeight parms (81ff6de)
• bump dependencies version (84058bb)

Release 1.7.0

FaceDetector now supports detection on any orientation following VisionCamera frame orientation in both front and back cameras. 🎉

• ⚠️ Minimum VisionCamera's required version is 4.1.0

✨ Features

• toggle camera (1ed399c)
• ios orientation (5a18291)
• android orientation - fix #62 (maybe #69 too) (d889e79)
• bump dependencies versions (a5f3d0b)
• rotating detection box (ef088b2)

🐛 Bug Fixes

• fix missing hook dependency (68f38b7)
• fix missing Worklets import (abfde7f)
• kotlin buildscript (dc9519e)

Release 1.6.3

• fix #56 orientation issue on IOS (185acf9)
• fix example (a1f51d7)

Release 1.6.2

• update tested versions (4abdc70)
• add expo privacyManifests (257b7c1)
• bump example app dependencies version (6184921)
• fix importings (b03412a)
• bump dependencies versions (689b42a)
• rename github profile (1c4e1bd)
• removed unnecessary comment (446b20e)
• runAsync using default context (02c6194)
• bump dependencies to latest version (1760f77)
• close #51 (8a5664d)
• fix camera permission text not in english (48a98de)

⚠️ BREAKING CHANGES ⚠️
In this new release we can now use latest vision-camera and worklets-core versions (see README's versions). Using default runAsync method causes the app to crash due to a worklets-core context creation issue but using defaultContext seems to work fine (see this issue).

It's not a definitive fix but it's working.

Release 1.6.1

• fix get frame orientation (ef2d546)
• lint (512b27b)
• workaround to #43 (5a0d896)
• downgrading vision-camera and worklets-core versions (f6d2bd2)
• adding worklet decorator to finally function (f555c39)
• using useRunInJS hook (e92da36)

⚠️ BREAKING CHANGES ⚠️

If you're not using the recommended way in the README you probably don't need to upgrade to this version.
If you're using vision-camera's 4.0.1 version or above you should use1.6.0 version (but maybe bellow issue happen to you).

Unfortunatelly there's an issue with react-native-worklets-core that makes a released (store published, APK, IPA, running expo on production mode, etc...) app crash in both Android and IOS.
To handle this I created a workaround on 5a0d896 but this temporary fix require you to use two specific versions:

• react-native-vision-camera must be at 4.0.0-beta.11 version
• react-native-worklets-core must be at 0.4.0 version

Double check in your package.json if dependencies contain any other characters besides the version (E.g: "react-native-workles-core": "^0.4.0"). If they do, remove them, leaving only the dependency's version. (E.g: "react-native-workles-core": "0.4.0").

Release 1.6.0

• Merge pull request #36 from nonam4/v1.6 (d0bf32b)
• Merge branch 'main' into v1.6 (94a3a91)
• using frame orientation (24034b6)
• fix wrong minFaceSize in docs (78ffc14)
• fix get frame rotation degrees (ffcd398)
• updating docs (75a668c)
• removed unused type (6587e4c)
• fix vision camera v4 changes (8522910)
• rollback 92bc1eb - don't return frame from native side (c2a0c4b)
• removing custom code + fix frame freezing (f676906)
• using latest vision camera version (6a10dce)
• fix scale declaration (2cea5b5)
• fix wrong variable name (aedb7ee)
• bump min ios version (085e83e)
• fix callback not override any method from its superclass (fc1bf5d)
• improve example app (cd611ec)
• update readme (5cc4b04)
• simplifying returns and usage (1f8b7da)
• returning frame from native side (92bc1eb)
• add auto scale option (fb51707)
• fix #30 (26e1da6)
• bump versions (f6aff5f)
• drop base64 frame conversion support (337d131)
• fix wrong example (22c87fe)

⚠️ BREAKING CHANGES ⚠️

• Frame base64 conversion support dropped: This feature doesn't make sense as this is a face detector plugin not a frame conversor.
• Native side returning only detected faces: The plugin will only return detected faces array from native side but you still have access to current frame.
• Auto scaling on native side: You can choose if you want the plugin to handle scaling or disable this feature and handle scaling by yourself on JS side. This also makes this plugin usable with Skia Frame Processor.

Release 1.5.0

• Passing face detection options to initFrameProcessorPlugin - thanks to @mrousavy (23a9138)
• Fix faces array type - thanks to @mrousavy (0f6305f)
• Bump tested versions (902cffa)

⚠️ BREAKING CHANGES ⚠️
You should now import useFaceDetector (instead of detectFaces) and pass detection settings directly to it. More details in README.

Release 1.4.2

• fix #23, fix #29 - thanks to @techmarajo (870e705)
• update release-it settings (8b8bd86)