Normalization of pictures for perception
I am preparing a service that simulates eye tracking in the image. For the sake of curiosity: an example of the normalization of the image under the foveal vision of a person:
The brighter the area, the more it is perceived by sight. Model - trichromacy (full color vision). Highlighting plots - logarithmic.
')
Well, the model for this normalized representation:
The main sections are shown correctly (by tests), but the model is filled up in the left corner: somewhere an algorithm error. We will fix it)
UPD
After a series of publications about eye-traking modeling , I have 8 targets with real heat maps from Toby Reader on which I test my algorithms. To be honest, I miss them. If there is an opportunity to conduct research on this issue, I will be glad.
The brighter the area, the more it is perceived by sight. Model - trichromacy (full color vision). Highlighting plots - logarithmic.
')
Well, the model for this normalized representation:
The main sections are shown correctly (by tests), but the model is filled up in the left corner: somewhere an algorithm error. We will fix it)
UPD
After a series of publications about eye-traking modeling , I have 8 targets with real heat maps from Toby Reader on which I test my algorithms. To be honest, I miss them. If there is an opportunity to conduct research on this issue, I will be glad.
Source: https://habr.com/ru/post/61750/
All Articles