/Team-Autonom_MBRDI

MBRDI 2020 Hackathon Project repository

Primary LanguagePython

Team-Autonom

Repository for work on Mercedes Benz Digital Challenge 2020 Hackathon

Usage (Deep Learning Version)

  1. Clone the Repository.
  2. To setup environment for running inference code. Use the requirements.txt provided in the repository.
$ pip3 install -r requirements.txt
  1. The custom model used is the cyclist-CNN_custom_architecture.model file.

Specification of model is discussed below.

  1. For performing inference use the predictor.py (for image) or prediction_video.py (for video)
  2. The model can be converted to lighter tflite model for use in constrained embedded devices.
$ python3 predictor.py

Usage (HoG + SVM approach)

  1. Clone the Repository.
  2. To setup environment for running inference code. Use the requirements.txt provided in the repository.
$ pip3 install -r requirements.txt
  1. Run the file hogcyclist.py <name of video/photo>.
$ python3 hogcyclist.py bicycle1.jpeg

Dataset Used

Tsinghua-Daimler Cyclist Detection Benchmark Dataset

Dataset Specifications

Architecture