/ai19_track2_hcmus

AI19_Track2_HCMUS

Primary LanguageJupyter Notebook

ai19_track2_hcmus

This is the repo for AI City Challenge, Track 2 by HCMUS Team

In this repo:

  • Code for triplet embeding networks, use it to embed both vehicle image & landmark parts
  • Code for feature/attribute classification networks, use it to classify pose & vehicle type
  • Code for extracting landmarks from veri datasets, use it to get dataset to train faster rcnn
  • Code for detecting landmark, faster rcnn code
  • Code for integrating individual module into pipeline, to produce final result.