ChaLearn Looking at People: Identity-preserved Human Detection (Track 2: Thermal) Forum

Go back to competition Back to thread list Post in this thread

> Label

Hi team,

Regarding coordinates..

x, center of the bounding box along horizontal axis of the image. Normalized to [0,1].
y, center of the bounding box along horizontal axis of the image. Normalized to [0,1].
width, width of the bounding box. Normalized to [0,1].
height, height of the bounding box. Normalized to [0,1].
The normalization of x, y, width, and height allows us to use the exact same labels for the images of Depth and Thermal despite having different image resolutions.

what do u mean normalization ?

this is image name vid00000_3QRMGtpS

0 0.719531 0.516667 0.195312 0.966667

can you tell me --> what value used to for normalizing the x-axis . ( x * 213 or 120)
can you tell me --> what value used to for normalizing the y-axis . ( y * 213 or 120)
can you tell me --> what value used to for normalizing the width . ( w * 213 or 120)
can you tell me --> what value used to for normalizing the height . ( h * 213 or 120)

When I tried to draw the rectangle on the thermal image with the above coordinates, it is drawing on wrong area.

thanks,
sabari

Posted by: sabarinathan @ Dec. 11, 2019, 8:48 a.m.

Dear sabari,

there was a typo there, it must have been:
"y, center of the bounding box along vertical axis of the image. Normalized to [0,1]."

And therefore:
x * 213, y * 120, width * 213, height * 120
should give you the bounding box that you can draw on top of the image.

Also I would suggest you to try with several images to know the calculation above is correct. Remember the position of the bounding box can be off, mostly depending on the speed of movement during acquisition. This won't be the case for test data, were synchronization is improved and positions are manually revised in individual modalities (not just in the color).

sabari

Posted by: aclapes @ Dec. 12, 2019, 7:19 a.m.

Thank you. I will check it.

Posted by: sabarinathan @ Dec. 15, 2019, 1:29 p.m.
Post in this thread