A salience map is a dynamic topographical map that integrates data from various feature maps to provide a real-number measure of conspicuity (salience). Initially, Koch and Ullman (1985) employed salience to predict location priority in visual search. However, because the characteristics that created the salience values are not stored in the map, calculations performed on the map contents were independent of the features that produced the salience values.
Salience maps have been proposed for computations other than search and processing priority, including attention-based motion perception (Lu and Sperling, 1995), isoluminant red-green grating motion (Lu, Lesmes, Sperling centroid computation (Sun, Chubb, Wright, Sperling, 2018), and frontal-plane distance perception (Gan, Sun, Sperling, 2021). Researchers show that primary forms, such as letters and numerals, may be identified not just when painted black or white but also when defined by edges, filled with different textures than the backdrop, or filled with colors that are isoluminant from the background.
Substance indifference implied that 2D shape perception is largely based on a salience map. They also show that brightness is not required for accurate form perception, isoluminant writing is easily legible, and so-called luminance artifacts are unimportant, implying that salience is sufficient for accurate 2D shape perception.
Reference:jov.arvojournals.org/article.aspx?articleid=2778465
