🖼️ Image Segmentation Application Using Segment Anything Model (SAM)

🔎 About

This project focuses on image segmentation using the Segment Anything Model (SAM), with points and bounding boxes as prompt encoders.

Project languages:

• en

📦 Dependencies

Name	Version
streamlit	1.32.2
streamlit-plotly-events	0.0.6
numpy	1.26.4
opencv-python	4.12.0.88
pillow	10.4.0
torch	2.8.0+cu126
plotly	6.3.0
segment_anything	1.0
matplotlib	3.10.5
onnx	1.18.0
pycocotools	2.0.10
onnxruntime	1.22.1

🖥️ Requirements

• Operating System (OS): Windows 10, Mac, Linux.
• python>=3.8, pytorch>=1.7 and torchvision>=0.8
• Integrated Development Environment (IDE): VSCode.
• Web Browser: Google Chrome, Microsoft Edge, Firefox, Safari.

⬇️ Installation

Make a directory

mkdir img_segmentation_app

cd img_segmentation_app

Create and activate environment

python -m venv venv

venv\Scripts\activate 

Install dependencies

pip install --upgrade typing-extensions

---

for CUDA compute platform:

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu126

for CPU only:

pip3 install torch torchvision

---

pip install git+https://github.com/facebookresearch/segment-anything.git

pip install streamlit==1.32.2 streamlit-drawable-canvas opencv-python pillow numpy matplotlib pycocotools onnxruntime onnx 

Download model

wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_b_01ec64.pth -O sam_vit_b.pth

Clone Repository

git clone https://github.com/wicaksonohanif/image_segmentation_project.git

Run App.

streamlit run app4.py

🥼 Author(s) / Contributor(s)

• Wicaksono Hanif Supriyanto

📚 References

• Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W.-Y., Dollár, P., & Girshick, R. (2023). Segment Anything. arXiv. https://arxiv.org/abs/2304.02643