ceng-407-408-2024-2025-Liewave-Lie-Detection-Using-Speech-Processing-Techniques

LieWave: Lie Detection Using Speech Processing Techniques

Project Description

LieWave is a mobile app that uses advanced machine learning models to detect lies by analyzing speech patterns. It processes audio inputs to identify subtle cues in tone, pace, and linguistic features, providing real-time insights with a lie probability score.

---

Team Members

Name	Student Number
Onur Güven	201611028
Eray Yıkar	202011074
Berkay Üğe	202011055
Çağrı Başaran	202011015
Melih Taşkın	202111070

---

Advisor

Prof. Dr. Hayri Sever

---

Installation Guide

1. Clone the Repository

git clone https://github.com/your-repo/ceng-407-408-2024-2025-Liewave-Lie-Detection-Using-Speech-Processing-Techniques.git
cd ceng-407-408-2024-2025-Liewave-Lie-Detection-Using-Speech-Processing-Techniques

2. Setup the Flask API Backend

2.1 Download Model Files

• Download the pre-trained machine learning model files from the following link:
  Google Drive - LieWave Model Files

• Place the entire model folder inside the liewave-API directory in your cloned repository.

---

2.2 Environment Variables

Create a .env file inside the liewave-API directory. This file will contain configuration values required by the Flask backend.

You need to set the following variables in the .env file:

• FLASK_APP
  Specifies the entry point script for Flask to run your application. Set to app.py.

• FLASK_ENV
  Defines the environment mode for Flask. Use development for debugging purposes.

• FIREBASE_CREDENTIALS_PATH
  The path to your Firebase Admin SDK JSON credentials file, which enables secure server-side access to Firebase services. You must download this JSON file from your Firebase project console and place it inside the API folder.

• STORAGE_BUCKET
  Your Firebase Storage bucket URL, which is where audio recordings and other files will be uploaded.

• MODEL_PATH
  Relative path to the folder containing your trained ML models. Should be set to model if you placed the model folder as instructed.

• FIREBASE_DATABASE_ID
  The Firestore database identifier for your Firebase project. Usually default unless configured otherwise.

• FFMPEG_PATH
  The full system path to the ffmpeg.exe executable, used for audio processing tasks. You must install FFmpeg on your machine and specify its path here.

---

2.3 Firebase Setup

Make sure your Firebase project is configured with the following services:

• Firebase Storage: For storing audio recordings.
• Firestore Database: For storing prediction metadata such as user ID, timestamp, prediction, and confidence.
• Firebase Authentication: For secure user authentication, especially Google Sign-In.

---

2.4 Install Python Dependencies

Before running the Flask API, install the required Python packages using:

pip install -r requirements.txt

2.5 Running the Flask API

After completing setup, navigate to the liewave-API directory and start the Flask API by running:

flask run

By default, the API will serve on http://localhost:5000 unless configured otherwise.

3. Setup the Flutter Mobile App

3.1 Firebase Configuration

Follow Firebase's official Flutter integration guide to set up Firebase in your Flutter project, including adding the necessary config files (google-services.json for Android, GoogleService-Info.plist for iOS).

---

3.2 Environment Variables for Flutter

Create a .env file in the root directory of the Flutter project containing:

• API_URL
  The base URL of the Flask API endpoint for prediction requests. This should include the IP address and port where your Flask API is running, for example:

API_URL=http://192.168.2.206:5000/predict

Make sure this URL is reachable from the device running the Flutter app.

3.3 Install Dependencies
From your Flutter project root, run the following commands to refresh dependencies and clean build cache:

flutter pub clean
flutter pub get

3.4 Running the Flutter App

Connect your device or start an emulator, then run:

flutter run

Important Notes and Requirements

• Internet Access: Both backend and mobile app require internet connectivity to communicate with Firebase services and the Flask API.

• Microphone Permission: The app requests microphone permission at runtime for recording audio. Ensure to grant this permission.

• Audio Format: The system records audio in AAC format and uploads it to Firebase Storage.

• FFmpeg Installation: For the backend audio processing, FFmpeg must be installed and the path correctly set in the .env file. FFmpeg is critical for audio file handling and preprocessing.

• Firebase Admin Credentials: Download from Firebase Console under Project Settings > Service Accounts. This file enables the backend to interact with Firebase securely.

• IP Address: If testing on a local network, ensure that your mobile device can access the Flask API server's IP address and port.

Troubleshooting Tips

Problem	Possible Cause	Suggested Solution
App cannot access microphone	Permission denied	Check device settings and allow microphone access
API requests fail	Incorrect API URL or network issue	Verify API_URL in Flutter .env and network connection
Audio not uploading	Firebase Storage misconfiguration	Check Firebase Storage rules and credentials
Model predictions missing	API or model errors	Check Flask API logs and ensure model is loaded correctly
Flutter build errors	Dependency or configuration issues	Run flutter clean and flutter pub get again

Additional Resources

• Flutter Firebase Setup
• FFmpeg Official Download
• Firebase Admin SDK
• Flask Documentation