https://github.com/haturusinghe/cnn-age-gender-detection-android

An Android application that uses TensorFlow Lite models (CNN) to detect age and gender from images using the device camera (realtime) or gallery images.
https://github.com/haturusinghe/cnn-age-gender-detection-android

cnn convolutional-neural-networks kotlin-android tensorflow tensorflow-lite

Last synced: 4 months ago
JSON representation

An Android application that uses TensorFlow Lite models (CNN) to detect age and gender from images using the device camera (realtime) or gallery images.

• Host: GitHub
• URL: https://github.com/haturusinghe/cnn-age-gender-detection-android
• Owner: haturusinghe
• Created: 2021-10-25T10:22:47.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2025-01-09T03:51:30.000Z (over 1 year ago)
• Last Synced: 2025-04-14T08:07:04.268Z (about 1 year ago)
• Topics: cnn, convolutional-neural-networks, kotlin-android, tensorflow, tensorflow-lite
• Language: Kotlin
• Homepage:
• Size: 3.56 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# TensorFlow Lite Age and Gender Detection Android App

An Android application that uses TensorFlow Lite models to detect age and gender from images using the device camera or gallery images.

## Features

- Age and gender detection from:

  - Device camera captures

  - Gallery image selection

- Multiple acceleration options:

  - CPU inference

  - GPU acceleration support

  - NNAPI support for compatible devices

- Real-time inference time reporting

- User-friendly interface with:

  - Camera capture button

  - Gallery selection button

  - Model initialization controls

  - Hardware acceleration toggles

## Technical Stack

- Language: Kotlin

- ML Framework: TensorFlow Lite

- Architecture Components:

  - Coroutines for asynchronous operations

  - View Binding for UI interactions

  - TFLite delegates (GPU, NNAPI) for acceleration

## Models

Refer to my [CNN Age Gender Detection](https://github.com/haturusinghe/cnn-age-gender) repo for model implementations

The application uses two TensorFlow Lite models:

- Age Estimation Model (`model_age_vN_nonq.tflite`)

  - Input size: 200x200 pixels

  - Output: Normalized age value

- Gender Classification Model (`model_gender_nonq.tflite`)

  - Input size: 128x128 pixels

  - Output: Binary classification [Male, Female]

## Requirements

- Android Studio Arctic Fox or newer

- Minimum SDK: Android 6.0 (API level 23)

- Android device with:

  - Camera

  - Storage access

  - (Optional) GPU/NNAPI support for acceleration

## Setup and Installation

1. Clone the repository

2. Open in Android Studio

3. Sync Gradle dependencies

4. Build and run on your device

## Usage Instructions

1. Launch the app

2. Initialize the models:

   - Select desired acceleration options (GPU/NNAPI if available)

   - Click "Initialize Models" button

3. After initialization, use either:

   - Camera button to take a photo

   - Gallery button to select an image

4. View results:

   - Predicted age

   - Predicted gender

   - Inference time for both models

## Permissions

- `CAMERA`: For capturing photos

- `READ_EXTERNAL_STORAGE`: For gallery image selection

## Project Structure

Key files:

- `MainActivity.kt`: Main UI and camera/gallery handling

- `AgeEstimationModel.kt`: Age prediction model wrapper

- `GenderClassificationModel.kt`: Gender prediction model wrapper

## License

This project is licensed under the MIT License.