A Flutter-based mobile application for object removal from images using on-device deep learning models.
The app provides a simple interactive workflow: select an image, draw a mask, and let the inpainting model reconstruct the missing region. All processing runs locally on the device using ONNX Runtime.

• Image selection
  Choose images from the gallery or local filesystem or take a photo.

• Automatic segmentation with MobileSAM (ONNX)
  Tap on the image or draw circle over the object to generate segmentation masks using MobileSAM, enabling object-aware removal workflows.

• On-device inpainting with MI-GAN (ONNX)
  The MI-GAN model performs high-quality inpainting with no internet connection required.

• Developer-friendly logging
  Logs inference times, model events, failures, and internal states to help with debugging and performance profiling.

• Flutter / Dart – cross-platform UI
• onnxruntime_flutter – optimized on-device ONNX inference
• MI-GAN ONNX – generative inpainting model
• MobileSAM ONNX – lightweight segmentation model for point-based prompting
• Target platforms: Android and iOS

• lib/app.dart – global app configuration (theme, navigation)
• lib/main.dart – app entry point
• lib/inpainting_page.dart – complete inpainting workflow UI
• lib/mask_painter.dart – canvas overlay for mask drawing
• lib/image_service.dart – image loading, resizing, saving
• lib/inpainting_service.dart – MI-GAN inference logic
• lib/segmentation_service.dart – MobileSAM encoder/decoder inference
• lib/image_utils.dart, lib/tensor_utils.dart – utilities for converting images to tensors and back
• lib/app_logger.dart – structured logging for debugging and performance metrics

The app is designed as a modular foundation for experimenting with mobile inpainting workflows.
By combining segmentation (MobileSAM) with generative inpainting (MI-GAN), the application aims to support advanced object-removal use cases entirely offline. The architecture allows for easy extension — such as integrating new models or benchmarking models.