---
title: Sidewalk Segmentation with SAM - Shiny Web App
emoji: 🌍
colorFrom: green
colorTo: indigo
sdk: docker
pinned: false
license: mit
---

# CS370 - Project Web App - Aaron Schwarzwaelder

This web app is a simple interface for running the sidewalk segmentation model on a user-provided image.
It is built using the Shiny framework in Python.
For more information on how to train the segmentation model refer to [this](https://github.com/aarons20/cs370-assignments/tree/main/project_1) repository.

## Development Environment Setup

To develop the web app locally you can use a virtual environment.
Install all required Python packages listed in `requirements.txt`.

To run the application locally change to the `app` directory and execute the following command.

```bash
shiny run --reload --launch-browser app.py
```

## Build the Container

The project uses a container that contains all required Python packages listed in `requirements.txt`.

To ensure a smooth startup of the container, follow the steps.
Make sure to execute the commands in the projects directory.

```bash
docker build -t project-web-app-image .
```

## Start the Container

```bash
docker run -p 8080:7860 project-web-app-image
```

## Publish on Hugging Face

After the development of the web app, the container can be published on Hugging Face.
To automate this process, this repository contains a GitHub Action that automatically publishes the container on Hugging Face.

The hosted application can be found on Hugging Face at the following link: [Sidewalk Segmentation with SAM](https://huggingface.co/spaces/UpToDataScience/SidewalkSegmentation_Shiny)

## Project Video

[This](https://www.youtube.com/watch?v=e8YOMNLTm0g) video provides short insights into the project.
It covers the main features and a demonstration of how the developed application works.