This repository contains the source code and resources for a Fire Detection Android application that utilizes the MobileNet model, optimized and compressed using TensorFlow Lite. The app's primary goal is to identify the presence of fire within video frames and provide real-time feedback to the user.
Detecting fires quickly and accurately is crucial for preventing and mitigating potential disasters. This Android app leverages machine learning and computer vision techniques to identify fires within video frames, providing valuable information to users in real-time.
The core of this application is the MobileNet model, which has been optimized and compressed using TensorFlow Lite. This ensures efficient performance on mobile devices without sacrificing accuracy. The app captures video frames from the device's camera and passes them through the model for inference.
- Android Studio
- TensorFlow Lite
MobileNet, a family of efficient convolutional neural network models, is an excellent choice for object detection tasks like fire detection in images or video frames. Its lightweight architecture, depthwise separable convolutions, and TensorFlow Lite compatibility make it well-suited for resource-constrained mobile and embedded devices. By fine-tuning a MobileNet model on fire-related data, it can quickly and accurately identify fires, enabling real-time fire detection applications that play a crucial role in enhancing safety and preventing disasters.
- Clone the GitHub Repository and run the
FireDetection.ipynbnotebook to generate the tensorflow lite model.
git clone https://github.com/aayushsss1/Fire-Detection.git- Open the project in Android Studio.
- Build and run the app on your Android device or emulator.
The Final Android Application !
