/light_obj_detection

MPhil thesis code

Primary LanguageJupyter NotebookMIT LicenseMIT

light_obj_detection

MPhil thesis code

Prior to runnint the detectors, follow the next steps:

  1. Create a data folder in the current repo folder
  2. Create a link to the Visdrone dataset, named VISDRONE
  3. build SolotNet/backbones/DCNv2
  4. build SolotNet/external

Please see the bash_scripts/power_test.sh script for examples on how to run the evaluation.

Please refer to run_examples/ for script examples on how to run the networks

Please refer to SolotNet/config.py for a full list of available parameters.