AI Blitz #7: Completed #educational Weight: 15.0
2631
331
15
184

🕵️ Introduction

You’ve successfully fixed the crisis and have taken off, phew! 🚀

But hey, what is that?! You’re encountering a large belt of space debris! You must use the tools of object detection to box this debris so you can safely take your rover to Mars.

💾 Dataset

The given dataset contains images of space with space debris. The images are of size 256*256 in jpg format. The bounding boxes are in bboxes with the columns as ImageId and bboxes0  containing list in [xmin, xmax, ymin, ymax] format.

A sample row :

 ImageID bboxes 0 [[34, 65, 69, 98], [144, 172, 266, 295], [382, 409, 248, 278], [383, 411, 438, 466]]

The boxes will be in the string but to convert them into a python list, you can simply use literal_eval function from ast python library!

📁 Files

Following files are available in the resources section:

• train.zip - (20000 samples) This zip file contains space with debris images, with images name corresponding to ImageID column of train.csv
• train.csv - (20000 samples) This csv file contains the ImageID column corresponding to train.zip and bboxes the column contains bounding boxes in a list.
• val.zip - (2000 samples) This zip file contains space with debris images, with images name corresponding to ImageID column of val.csv
• val.csv - (2000 samples) This csv file contains the ImageID column corresponding to val.zip and bboxes the column contains bounding boxes in a list.
• test.zip - (5000 samples) This zip file contains space with debris images which will be used to evaluate the performance of the model.

🚀 Submission

• Prepare a CSV containing ImageID column corresponding to test.zip and bboxes column as bounding boxes. Note that the bboxes column should contain another element in every bounding box list containing scores. Something like below:-
 ImageID bboxes 0 [[34, 65, 69, 98, 0.98], [144, 172, 266, 295, 0.8], [382, 409, 248, 278, 0.55], [383, 411, 438, 466, 0.23]]

• The name of the above file should be submission.csv.
• Sample submission format available at sample_submission.csv in the resources section.

Make your first submission here 🚀 !!

🖊 Evaluation Criteria

During the evaluation, Average Precision (AP) @[ IoU=0.50:0.50 | area=medium | maxDets=100 ] will be used to test the efficiency of the model.