MyDeeplabV3 is a Python implementation of my paper, "Semantic Segmentation for 3D Reconstruction of ASINs for AR". This project leverages the DeepLabV3 architecture for semantic segmentation to support 3D reconstruction of objects (ASINs) for augmented reality (AR) applications. It focuses on accurately segmenting objects in images to facilitate high-quality 3D model generation.

• Semantic Segmentation: Utilizes DeepLabV3 for precise object segmentation in images.
• 3D Reconstruction Support: Generates segmentation masks optimized for AR-ready 3D models.
• Customizable Model: Allows fine-tuning of DeepLabV3 for specific object categories.
• Visualization Tools: Includes utilities to visualize segmentation results.

1. Clone the repository:

   git clone https://github.com/xzhang311/MyDeeplabV3.git

2. Navigate to the project directory:

   cd MyDeeplabV3

3. Install dependencies:

   pip install -r requirements.txt

Run the main script to perform semantic segmentation on an input image:

python main.py --input path/to/image.jpg --output path/to/segmentation_mask.png

Use the --help flag for detailed configuration options, including model weights and output formats:

python main.py --help

• Python 3.8+
• Libraries: PyTorch, torchvision, NumPy, OpenCV, Matplotlib (listed in requirements.txt)

This project implements the methodology described in:

• "Semantic Segmentation for 3D Reconstruction of ASINs for AR".

Contributions are welcome! Please fork the repository, create a feature branch, and submit a pull request with your enhancements.

This project is licensed under the MIT License. See the LICENSE file for details.

This work is based on the DeepLabV3 architecture and research on semantic segmentation for 3D reconstruction in AR applications.