IMPORTANT: Before starting this challenge, make sure we have a position by sending your resume and a motivated email to [email protected].
Please use this link to download the source code and do not fork the repository. Please also read all the instructions carefully.
The purpose of this simple exercise is to do some basic bounding box manipulation, grouping, verifying, etc. For more information simply go into the 1_axle_tree_exercise directory and open the README.md file.
This time you are given both the model prediction and the ground truth for the images and you need to build a module that evaluates the performance of the network. Same thing, just go to 2_precision_accuracy/README.md.
Once you are happy with the results, please zip this folder and provide us with any information needed to run your code.
Please send an archive with your solution to your point of contact. You will be judged on:
- the quality of your code
- the elegance of your solution
- the method you adopted
If you think it's a good idea to develop additional scripts, for instance to let you draw boxes on the actual images in the images directory. Please feel free to include them in your submission!
Good luck ! :-)