search

dangweili / pedestrian-attribute-recognition-pytorch

A simple baseline for pedestrian attribute recognition in surveillance scenarios
332 stars 80 forks source link

the score on the demo_image_result.png #12

Open kclch opened 5 years ago

kclch commented 5 years ago

Hello, I chose a picture to test, but I don't understand the score on the demo_image_result.png. What does this score mean? Thank you very much image

dangweili commented 5 years ago

It is the classification score of the corresponding attribute before sigmoid function.

kclch commented 5 years ago

Thank you,Why not use the sigmoid function to process the scores? Because the scores before classification are not intuitive

At 2019-03-12 00:18:32, "Dangwei Li" notifications@github.com wrote:

It is the classification score of the corresponding attribute before sigmoid function.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

dangweili commented 5 years ago

This is just an example. I didn't pay much attention on it in the past. You can use probability if necessary.

ghost commented 5 years ago

@dangweili won't using the sigmoid be wrong? The attributes are independent and more than one attribute can have score 1 at the same time?