by Paul Curzon, Milan Verma and Hamit Soyel, Queen Mary University of London
This page accompanies the blog post “Spot the difference – modelling how humans see the world“.
Saliency Maps
The saliency map generated by the AI model of human vision shows where your eye is most likely to drawn to in red, with less likely areas in blue that your eye will slide away from.
“Salience” means “the quality of being particularly noticeable or important; prominence” (source Oxford Languages and Google).
Spot the Difference: Answer to Challenge 1
The tower in the context of the rest of the image represents an example of an area with low saliency. So this tends to be difficult to spot.

Spot the Difference: Answer to Challenge 2
The sign in the context of the rest of the image represents an example of an area with high saliency. So this tends to be easier to spot.

This blog is funded through EPSRC grant EP/W033615/1.