update README

CapMl.podspec

@@ -3,7 +3,7 @@
     s.name = 'CapMl'
     s.version = '0.0.1'
     s.summary = 'Machine Learning Plugin for Capacitor'
-    s.license = 'Apache v2'
+    s.license = 'Hippocratic License Version 2.0'
     s.homepage = '[email protected]:bendyworks/cap-ml'
     s.author = 'Vennela Kodali'
     s.source = { :git => '[email protected]:bendyworks/cap-ml', :tag => s.version.to_s }

README.md

@@ -3,7 +3,7 @@
 Machine Learning Plugin for Capacitor. Currently offered implementations include -
   - Text Detector: Text Detection in still images.
 
-    We're using Apple's Vision Framework which has some limitations like not being able to detect cursive/handwriting font etc.
+    On the ios side, we're using Apple's Vision Framework and MLKit's Vision Framework on the Android side. Both have some limitations like not being able to detect cursive/handwriting font etc.
 
     TextDetector expects the image to be sent in portrait mode only, i.e. with text facing up. It will try to process even otherwise, but note that it might result in gibberish.
 
@@ -81,6 +81,8 @@ After checking out the repo,
   - run `npx cap sync` to sync with ios and install pods
   Plugin should be ready at this point. To test it out -
   - navigate to examples/text-detection/ImageReader
+
+  ### ios Development
   - run `npx capacitor open ios` to open up an XCode project.
   - Run the XCode project either on a simulator or a device.
   - For each change in the javascript part of the app, run `npm run build && npx cap sync ios` to deploy the corresponding changes to ios app
@@ -89,6 +91,16 @@ After checking out the repo,
   Plugin code is located at Pods/DevelopmentPods/CapML
   - `Plugin.swift` is the entry point to the Plugin.
 
+  ### Android Development
+  - run `npx capacitor open android` to open up an Android Studio project.
+  - Run the project either on a simulator or a device.
+  - For each change in the javascript part of the app, run `npm run build && npx cap sync android` to deploy the corresponding changes to android app
+    (or)
+  - (recommended) Enable live reload of the app, using `ionic capacitor run android --livereload`
+  - Plugin code is located at `android-cap-ml/java/com.bendyworks.capML`
+  - `CapML.java` is the entry point to the Plugin.
+    (Note: When plugin code is updated, make sure to rebuild the project before running it.)
+
 ## Contributing
 
 Bug reports and pull requests are welcome on GitHub at https://github.com/bendyworks/cap-ml.

android/src/main/java/com/bendyworks/capML/TextDetector.kt

@@ -37,8 +37,7 @@ class TextDetector {
               val rect: Rect = line.boundingBox ?: throw NoSuchPropertyException("FirebaseVisionTextRecognizer.processImage: could not get bounding coordinates")
 
               val textDetection = mapOf(
-                // normalizing coordinates and switching to 1st quadrant ie origin(0,0) as bottom left
-                // in order to match the result with that of core-ml's Vision framework.
+                // normalizing coordinates
                 "topLeft" to listOf<Double?>((rect.left).toDouble()/width, (height - rect.top).toDouble()/height),
                 "topRight" to listOf<Double?>((rect.right).toDouble()/width, (height - rect.top).toDouble()/height),
                 "bottomLeft" to listOf<Double?>((rect.left).toDouble()/width, (height - rect.bottom).toDouble()/height),

examples/text-detection/ImageReader/src/app/home/home.page.ts

@@ -49,7 +49,7 @@ export class HomePage implements OnInit {
     this.textDetections.forEach((detection: TextDetection) => {
       // the received coordinates are normalized by Vision framework, proportionate to the width and height of the image itself. Hence here, x coordinate is multiplied by image width(scaleX), and y with height(scaleY) to obtain de-normalized coordinates for the chosen image scale.
       const x = detection.bottomLeft[0] * this.scaleX
-      // In addition to de-normalizing, subtracting from scaleY because TextDetector(which uses Apple's Vision Framework) assumes bottom-left as origin (0,0) vs SVG rect which assumes top-left as origin (0,0)
+      // In addition to de-normalizing, subtracting from scaleY because cap-ml assumes bottom-left as origin (0,0) vs SVG rect which assumes top-left as origin (0,0)
       const y = this.scaleY - (detection.topLeft[1] * this.scaleY)
 
       // Similar to the x and y coordinates above, the received coordinates are normalized by Vision framework, proportionate to the width and height of the image itself. Hence here, difference between corresponding x-coordinates is multiplied by image width(scaleX), and difference between corresponding y-coordinates is multiplied by image height(scaleY) to obtain de-normalized dimensions for the chosen image scale.

package.json

@@ -11,7 +11,7 @@
     "prepublishOnly": "npm run build"
   },
   "author": "Vennela Kodali",
-  "license": "Apache v2",
+  "license": "Hippocratic License Version 2.0",
   "dependencies": {
     "@capacitor/core": "latest"
   },