FacadeStream

Visualization tool for Facade Scanner Project

Hosted with Streamlit!

https://share.streamlit.io/kai98/facadestream/main/scanner.py

TL;DR

Installing Dependencies

In your virtual environment (for example, anaconda), install all dependencies by running:

pip3 install -r requirements.txt

Running the Streamlit App

streamlit run scanner.py

Then download the pretrained model by clicking the button.

How to Run the App

Step 0: Virutal Environment [Optional]

It is convenient to use pip to manage python environment.

To check the current pip version:

pip3 --version

Installing the latest pip:

pip3 install --upgrade pip

Installing virtualenv

virtualenv: https://pypi.org/project/virtualenv/

pip3 install virtualenv

Creating a virtual environment, 'facade-env' (or any other names)

virtualenv facade-env

Activating the virtual environment

Windows

facade-env\scripts\activate

Mac & Linux

source facade-env/bin/activate

Deactivating the virtual environment

To leave the virtual environment when everything is done (the localhost is running, in our case).

deactivate

Step 1: 'Pip Installs Packages'

Change directory to the folder. In your pip environment, install the dependencies using:

pip3 install -r requirements.txt

Step 2: Import Trained Model

Download the pretrained model by clicking the button.

Download the trained model using following link:

https://drive.google.com/uc?export=download&id=1rJ3edeARtcprrgs14lj5iZLTLkn9kufw

, and put it under /models folder.

Step 3: Run the scanner app

Streamlit: https://streamlit.io

Start the facade segmentation tool by running:

streamlit run scanner.py

Others

Color Map

Using eTRIMES's color code

More about eTRIMS dataset: http://www.ipb.uni-bonn.de/projects/etrims_db/

Index	Label	RGB
0	Various	(0, 0, 0)
1	Wall	(128, 0, 0)
2	Car	(128, 0, 128)
3	Door	(128, 128, 0)
4	Pavement	(128, 128, 128)
5	Road	(128, 64, 0)
6	Sky	(0, 128, 128)
7	Vegetation	(0, 128, 0)
8	Window	(0, 0, 128)

kai98/FacadeStream