(IOU) Intersection over union explained. With Python Code. IOU formula explained.

In this video, we will see the concept of Intersection Over Union and also try to code it in Python.
The video first explains all the types of positions of the ground truth box and the predicted bounding box. It then explains how the boxes are labelled and how how to calculate the area of the intersection and also the area of the union. The video also explains how to get the formula for the box co-ordinates. Lastly, the video also has a Python implementation of the IOU whose code is available on Github.
Code: github.com/Dat...

[@MmAaZzzz]

very well explained

[@tinyeinmoe5147]

Thank you

[@atishayjain2190]

Hey...i have been doing some project and i have some doubts regarding this IoU concept. It would be great if you could help me clarify them.

[@ildarkhalilov9595]

why y1 < y3 and y2 < y4?

[@mondherbenromdhane2069]

did you take the use case where the the blocks can be separated ?

[@kauffmanes8269]

If there's no overlap, I think you have to get the area intersection like:
area_intersection = max(0, xr-xl) * max(0, yr-yl) so the intersection is 0

[@ياسمينشلبي-ك1ع]

@Datum Learning how can i know the x and y for each box, how did you know the value of them?

[@aswina7666]

Hello your videos are pretty good .
I have a question on YOLO . Does YOLO always return 8 values ???

[@datumlearning6204]

what do you mean by 8 values?

[@deepakpawade4084]

​@@datumlearning6204 They probably saw your old video on YOLO architecture where you used 3 classification classes. So the vector for each unit grid was of 8 values.

[@archanasoman-electronicswo6062]

how to get IoUs for an image frame with many objects and then find out Tp,FP,FN