StreetScanner

The system leverages DeepLabv3 ResNet-50 to detect and segment pedestrians and vehicles, with color-coded masks for easy visualization. This solution is designed for applications in traffic monitoring, pedestrian safety and smart city solutions. Built using PyTorch, Semantic Segmentation and Computer Vision techniques.

Execution Guide:

1. Clone the repoistory:

   git clone https://github.com/kr1shnasomani/TraffiSense.git
   cd TraffiSense/StreetScanner
   

2. Download the dependencies:

   pip install -r requirements.txt
   

3. On running the code it will save the results in the file - both-resultant-image.png, pedestrian-resultant-image.png and vehicle-resultant-image.png

Model Prediction:

Input Image:

Output Image:

a. both-resultant-image.png

b. pedestrian-resultant-image.png

c. vehicle-resultant-image.png

• vehicle.py: For applications requiring vehicle detection (e.g., traffic monitoring, autonomous vehicle systems).

These scripts demonstrate a modular approach, where specific objects of interest (pedestrians, vehicles) can be segmented independently or collectively, depending on the application.