Video Interpolation Android App

An Android application that performs video interpolation to create smooth slow-motion videos by generating intermediate frames between existing frames.

Features

• 📹 Video Frame Extraction: Automatically extracts frames from selected videos
• 🤖 AI Frame Interpolation: Uses machine learning to generate smooth intermediate frames
• ⚡ Adjustable Frame Rate: Choose from 2x to 10x frame rate multiplier
• 📱 Modern UI: Clean Material Design interface
• 🎬 Video Encoding: Automatically encodes interpolated frames back into video

Technical Stack

• Language: Kotlin
• ML Framework: TensorFlow Lite
• Media Processing: MediaCodec, MediaMuxer, MediaMetadataRetriever
• UI: Material Design Components, ViewBinding
• Async: Coroutines

Project Structure

app/
├── src/main/
│   ├── java/com/videointerpolation/app/
│   │   ├── MainActivity.kt              # Main UI and orchestration
│   │   ├── ml/
│   │   │   └── FrameInterpolator.kt     # ML-based frame interpolation
│   │   └── utils/
│   │       ├── VideoProcessor.kt        # Frame extraction
│   │       └── VideoEncoder.kt          # Video encoding
│   ├── res/
│   │   ├── layout/
│   │   │   └── activity_main.xml        # Main UI layout
│   │   └── values/
│   │       ├── strings.xml
│   │       ├── colors.xml
│   │       └── themes.xml
│   └── AndroidManifest.xml
└── build.gradle

How It Works

1. Frame Extraction: The app extracts individual frames from the selected video
2. Interpolation: For each pair of consecutive frames, the ML model generates intermediate frames
3. Encoding: All frames (original + interpolated) are encoded into a new high-framerate video
4. Output: The resulting video has a higher frame rate, creating smooth slow-motion effect

Setup Instructions

Prerequisites

• Android Studio Arctic Fox or newer
• Android SDK 24 or higher
• Device or emulator running Android 7.0+

Build Steps

1. Open the project in Android Studio
2. Sync Gradle files
3. Build the project: Build > Make Project
4. Run on device or emulator

Adding ML Model (Optional Enhancement)

The current implementation uses linear interpolation. For better results, add a TensorFlow Lite model:

1. Download a frame interpolation model (e.g., FILM, RIFE)
2. Convert to TFLite format if needed
3. Place the .tflite file in app/src/main/assets/
4. Update FrameInterpolator.kt to load and use the model

Recommended models:

• FILM (Frame Interpolation for Large Motion)
• RIFE (Real-Time Intermediate Flow Estimation)
• DAIN (Depth-Aware Video Frame Interpolation)

Usage

1. Launch the app
2. Tap "Select Video" to choose a video from your device
3. Adjust the frame rate multiplier (2x-10x) using the slider
4. Tap "Process Video" to start interpolation
5. Wait for processing to complete
6. Find the output video in the app's files directory

Permissions

The app requires the following permissions:

• READ_EXTERNAL_STORAGE / READ_MEDIA_VIDEO: To access videos
• WRITE_EXTERNAL_STORAGE: To save processed videos (Android 9 and below)

Performance Considerations

• Processing Time: Varies based on video length, resolution, and device performance
• Memory Usage: Large videos may require significant memory
• Storage: Processed videos are saved to external storage
• Battery: Video processing is CPU/GPU intensive

Future Enhancements

• Add GPU acceleration for faster processing
• Implement background processing with notifications
• Add video trimming/cropping before interpolation

Troubleshooting

Issue: Video encoding fails

• Solution: Ensure sufficient storage space and valid video format

Issue: Out of memory errors

• Solution: Try processing shorter videos or reduce resolution

Issue: Slow processing

• Solution: Use lower frame rate multiplier or shorter videos

License

This project is provided as-is for educational purposes.

Credits

Built with Android Studio and TensorFlow Lite