CarDetectorApp 🚙

This app was created in research goals. How does this work?:

• Open webpage of this app
• Upload street with cars image
• Wait for result (see Demo section)

Kaggle research and used model (implemented SegNet architecture) training process: link

Tech Stack

Client: HTML, CSS, JS

Server: Fastapi (Python 3.10), uvicorn

Features

CNN for this task was trained with lack of data (train dataset ~2500 segmented images). The expirement was to improve segmentation result with some "tricks". Main feature:

Run Locally

Clone the project:

  git clone https://github.com/ivanshin/CarDetectorApp.git

Switch to project directory:

  cd CarDetectorApp

Create virtual enviroment and activate:

  python -m venv .venv

  . .venv\Scripts\activate

Install requirements:

  pip install -r requirements.txt

Start uvicorn server:

  uvicorn app.main:app

Check if it works:

Visit: 127.0.0.1:8000

Hugging face

---

title: Car Detector emoji: 🚙 colorFrom: white colorTo: blue sdk: docker pinned: true

Demo