MultiScan is an Android application that combines document scanning and landmark recognition capabilities. Built with Jetpack Compose for a modern UI and TensorFlow Lite for landmark recognition, this project showcases advanced Android development skills.

• Document Scanner: Scan documents and save them as PDF files
• Landmark Recognition: Recognize famous landmarks using TensorFlow Lite models

This project has been set up with the basic integration structure, but you need to complete the following steps to fully enable the landmark recognition functionality:

1. Copy TensorFlow Lite models from Landmark-Recognition2 project:

   Copy the following .tflite model files from Landmark-Recognition2/app/src/main/assets to MultiScan/app/src/main/assets:

   • africa.tflite
   • asia.tflite
   • europe.tflite
   • northamerica.tflite
   • oceania_antarctica.tflite
   • southamerica.tflite

2. Copy needed classifier implementation classes:

   Copy the following files from Landmark-Recognition2/app/src/main/java/com/elsharif/landmarkrecognition/data to similar package structure in MultiScan:

   • TFLiteLandmarkClassifierOfAfrica.kt
   • TFLiteLandmarkClassifierOfAsia.kt
   • TFLiteLandmarkClassifierOfEurope.kt
   • TFLiteLandmarkClassifierOfNorthAmerica.kt
   • TFLiteLandmarkClassifierOfOceaniaAntarctica.kt
   • TFLiteLandmarkClassifierOfSouthAmerica.kt

3. Copy domain classes:

   Copy the domain classes from Landmark-Recognition2/app/src/main/java/com/elsharif/landmarkrecognition/domain to your project.

4. Enhance the LandmarkScreen implementation:

   Update the basic LandmarkScreen.kt to use the classifiers and implement camera functionality similar to the Landmark-Recognition2 project.

• app/src/main/java/com/example/multi_scan/navigation: Contains navigation-related code
• app/src/main/java/com/example/multi_scan/ui/screens/home: Contains the home screen and document scanner screen
• app/src/main/java/com/example/multi_scan/ui/screens/landmark: Contains the landmark recognition screen
• app/src/main/assets: Should contain the TensorFlow Lite models

This project uses Gradle to build. Open the project in Android Studio and run as a regular Android application.

The welcome screen with options for Document Scanner and Landmark Recognition.

• The home screen also has the option to change the app language, which supports up to 4 languages.

Selecting a region (e.g., Asian Landmarks) for recognition.

Recognizing India Gate with 89% confidence using TensorFlow Lite.

Document scanner showing saved PDFs (e.g., baka.pdf, sussy.pdf).