Feature : Image Classifier

[narshi]

Description :

feat: Add on-demand MNN image classification from gallery

Integrate MNN-based image classification using a pre-copied MobileNet-v1 model, allowing users to analyze images picked from the device gallery.

Functional Changes

• MNN Integration: Implemented runClassifier() to handle model loading, image preprocessing (resize/normalize to NCHW [1, 3, 224, 224]), and calling the native MNNClassifier.run() wrapper.
• UI Trigger: Added the 'Pick Image' button (btnPickImage) to activity_main.xml and hooked it up in MainActivity to launch the image picker via startActivityForResult.
• Pipeline: Handled the result in onActivityResult to trigger the runClassifier method.
• Result Display: Displays the predicted class index using a Toast message upon successful inference.

Technical Details

• Ensured necessary model copying logic from assets to the filesystem during the initial classifier button setup.
• Implemented bitmapToNCHWFloatArray helper for required input preprocessing.
• Used lifecycleScope and Dispatchers.IO for running the inference task off the main thread.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Narayanks seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}