abebaw8760/Currency-Detector-For-Blind

This is mobile app that is used to detect the currency for blind, the app use different tech kotlin, jetpack compose use camerax, and also the app us tensorflow file that is modeled with Ethiopian Currecny # Ethiopian Currency Detector for Blind Users An Android app that automatically detects Ethiopian banknotes (5, 10, 50, 100, 200 birr) using the camera and announces the denomination with a female voice. Designed specifically for accessibility and blind users. ## Features ✨ **Real-time Detection**: Uses CameraX for live camera feed analysis 🧠 **AI-Powered**: TensorFlow Lite MobileNetV2 for fast, accurate inference 🔊 **Voice Announcements**: Text-to-Speech with female voice preference ♿ **Accessibility**: Full TalkBack support and large touch targets 🎯 **Smart Detection**: Debouncing to avoid false announcements 🔇 **Mute Control**: Toggle audio on/off as needed 📱 **Offline**: Works completely offline, no internet required ## Screenshots ``` ┌─────────────────────────┐ │ │ │ CAMERA PREVIEW │ │ │ │ │ │ 📷 │ │ │ │ │ └─────────────────────────┘ ┌─────────────────────────┐ │ Status: Detecting... │ │ │ │ 50 Birr │ │ 87.3% │ │ │ │ 🔊 Mute Audio [OFF] │ └─────────────────────────┘ ``` ## Requirements - **Android Device**: Android 5.0 (API 21) or higher - **Camera**: Back-facing camera required - **Model File**: `ethiopia_currency_mobilenetv2.tflite` (not included) - **Storage**: ~50MB free space ## Setup Instructions ### 1. Clone/Download the Project ```bash git clone cd CurrencyDetector ``` Or extract the ZIP file if downloaded. ### 2. Open in Android Studio 1. Open Android Studio (Recommended: Arctic Fox or newer) 2. Click **File > Open** and select the `CurrencyDetector` folder 3. Wait for Gradle sync to complete ### 3. Add the TensorFlow Lite Model **CRITICAL STEP**: You must provide your own trained model. 1. Train or obtain a MobileNetV2 TFLite model for Ethiopian currency detection 2. Name the file exactly: `ethiopia_cu …