BharatScan

Features | Screenshots | Quick Start | Tech Stack | License

BharatScan is a Made in India Android document scanning app built with Jetpack Compose, CameraX, and on-device ML for segmentation and OCR. This repo includes the Android app plus supporting JVM modules for image processing and evaluation.

Version Updates

v1.1.0 (Current)

• Tricolor Modern UI refresh with improved visual separation and India-inspired styling.
• Background image support for Home, Documents, and Settings with tint overlays.
• OCR Search Index: search inside scanned text across documents.
• Guided walkthrough across Home → Scan → Edit → Export → Search.
• Update flow polish with GitHub Releases checks, progress UI, and formatted notes.
• Scanner and crop refinements (retake per page, manual crop fallback).

Features

• Camera-based document capture and scanning workflow.
• Automatic document boundary detection and auto-cropping via an on-device segmentation model.
• PDF export with optional OCR text layer (ML Kit Text Recognition).
• Built-in PDF viewer with zoom and search highlights.
• PDF password handling and optional security gate via device biometrics.
• Export formats and quality controls (see settings and export screens in the app).
• External intent support for org.bharatscan.app.action.SCAN_TO_PDF and PDF VIEW/EDIT intents.

Screenshots

Home	Document	Camera
Export	Success	PDF Viewer
Settings	Settings Detail

Tech Stack

• Kotlin 2.3.10 and Java 11 bytecode.
• Android Gradle Plugin 8.13.2.
• Jetpack Compose (BOM 2026.03.00).
• CameraX, ML Kit Text Recognition, PDFBox-Android.
• LiteRT (TFLite runtime) for on-device segmentation.
• OpenCV (JVM bindings) in imageprocessing.

Project Structure

• app Android application module (Compose UI, CameraX, PDF, OCR, settings, export).
• imageprocessing JVM library with OpenCV-backed image utilities and transforms.
• evaluation JVM module for evaluation workflows built on imageprocessing.
• gradle Version catalog and Gradle plugin configuration.
• metadata Play/metadata assets (if used for releases).

Requirements

• JDK 11.
• Android Studio (recommended) or command-line Gradle.
• Android SDK with API 36 installed.
• Network access during first build to download the segmentation model.

Quick Start

The project uses the Gradle wrapper.

# Debug build
./gradlew :app:assembleDebug

# Install to a connected device
./gradlew :app:installDebug

Windows PowerShell:

.\gradlew.bat :app:assembleDebug
.\gradlew.bat :app:installDebug

Model Download

The segmentation model is downloaded during preBuild.

• Task: downloadTFLiteModel in app/download-tflite.gradle.kts.
• Model name: fairscan-segmentation-model.tflite.
• Download location: app/build/downloads/ and copied into generated assets.

If you are building offline, run once with network access or pre-seed the model file into app/build/downloads/.

Build Configuration

• minSdk: 26
• targetSdk: 36
• compileSdk: 36
• ABI splits are enabled for APKs. Output names follow BharatScan-<versionName>-<abi>.apk.

Signing

Release signing is enabled when these Gradle properties are present:

• RELEASE_STORE_FILE
• RELEASE_STORE_PASSWORD
• RELEASE_KEY_ALIAS
• RELEASE_KEY_PASSWORD

Without these, the release build type will still be configured but unsigned.

Permissions

The app uses:

• android.permission.CAMERA
• android.permission.WRITE_EXTERNAL_STORAGE (Android 9 and below, for saving files)

Intents

MainActivity handles:

• Launcher entry point.
• Custom action: org.bharatscan.app.action.SCAN_TO_PDF.
• PDF view/edit intents for application/pdf with file and content schemes.

Testing

Unit and instrumentation tests are configured in the app module.

./gradlew :app:test
./gradlew :app:connectedAndroidTest

Contributing

Issues and PRs are welcome. Please keep changes small and focused, and include test updates where appropriate. See CONTRIBUTING.md for details.

License

This project is licensed under the GNU General Public License, version 3 or later (GPL-3.0-or-later).