---
title: Crosswalk Detection App
emoji: ⚡
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 5.7.1
app_file: app.py
pinned: false
thumbnail: >-
  https://cdn-uploads.huggingface.co/production/uploads/674d3ccd6b4dbbd9a8561126/Ewwg1TM1nsc3pkqtHI2Cb.png
---
Check out my first app on Hugging Face Spaces - https://lnkd.in/gyHSTTN6.

What does it do? Detects crosswalks in images of streets/intersections. Simply upload an image and get segmentation masks for crosswalks vs. rest of image, along with % pixels covered by crosswalks in the image. 

Why does it matter? Object detection models often struggle with out-of-distribution objects such as crosswalks. Whereas image segmentation makes it possible to create masks for any object. 
Using the % pixels covered in the image, one can (theoretically) detect crosswalks if they cover > X% of the image. Ref recent research from U.S. Department of Transportation here - https://lnkd.in/gX7XTU3v. 

Model training and app development? I trained a custom segmentation model on LandingAI, and used Gradio to write the program - it was easy to publish it in Spaces thereafter. 

I hope you try it out and share your feedback with me! 

Sample outputs ->
https://cdn-uploads.huggingface.co/production/uploads/674d3ccd6b4dbbd9a8561126/Ewwg1TM1nsc3pkqtHI2Cb.png
https://cdn-uploads.huggingface.co/production/uploads/674d3ccd6b4dbbd9a8561126/feS_f7WzB8PPotu7zLFd_.png


Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference