YOLO11s Android App

This is an Android application (Java) that runs the YOLO11s object detection model on mobile devices using TensorFlow Lite (TFLite). The app allows object detection on captured frames using the YOLO11s model optimized for edge devices.

Features

• Runs YOLO11s model on Android devices using TFLite.
• Live camera feed with manual detection.
• Captures the current frame when the Detect button is clicked.
• Displays detected objects with bounding boxes, class names, and confidence scores.

Model Conversion

To convert a YOLO model into the TFLite format, refer to the YOLO11S_Convert_to_TFLite.ipynb notebook included in this repository.

Pretrained Model

This repository includes a YOLO11s model pretrained for playing card symbol detection. However, you can use your own trained model by converting it to TFLite format and replacing the provided model.

Screenshots

Detection Output

Tensor Shape Output

The output tensor shape and its details (for COCO dataset, 80 classes) can be found in the following GitHub discussion: Ultralytics Discussion #17254

Requirements

• Android 7.0 (API level 24) or higher
• TensorFlow Lite dependencies added to build.gradle
• A device with a camera for capturing frames

Setup & Installation

1. Clone this repository:

   git clone https://github.com/tharushaudana/YOLO11S-TFLite-Android-Java.git

2. Open the project in Android Studio.
3. Ensure that TensorFlow Lite dependencies are added in build.gradle:

   implementation("org.tensorflow:tensorflow-lite:2.9.0")
   implementation("org.tensorflow:tensorflow-lite-task-vision:0.3.1")
   implementation("org.tensorflow:tensorflow-lite-gpu:2.9.0")

4. Place the TFLite model and classes.txt inside the assets folder.
5. Run the app on an Android device/emulator.

Usage

1. Launch the app.
2. Grant camera permissions.
3. View the live camera feed.
4. Click the Detect button to capture the current frame.
5. The detected objects will be displayed with bounding boxes, class names, and confidence scores.

Acknowledgments

• TensorFlow Lite
• Ultralytics