Back

More is better: Improved visual discrimination through local feature diversity in naturalistic objects

Glukhova, M.; Crider, A.; Galiagina, A.; Mert, B.; Shapcott, K.; Weigand, M.; Havenith, M. N.; Schölvinck, M. L.

2026-01-12 neuroscience
10.64898/2026.01.11.698882 bioRxiv
Show abstract

Visual discrimination is traditionally studied with highly controlled, artificial stimuli such as gratings. Yet real-world object discrimination relies on many interacting image dimensions. Here we compared human discrimination performance of oriented gratings with that of natural object morphs, and then asked which stimulus dimensions are most important for morph discrimination. In Experiment 1, participants discriminated between two artificial gratings, two natural grating-like textures, and two natural, morphed stimuli. Morph stimuli were discriminated faster and over a larger stimulus range. In Experiment 2, we separated the contribution of distinct image features, specifically colour, texture, and shape, to the discriminability of the morph stimuli. To this end, we varied two feature dimensions at a time while keeping the third dimension constant. Discrimination was best when colour and texture varied together, resulting in steeper psychometric functions and faster responses than when shape changed. Together, these results suggest a discriminative advantage for natural-object morphs over gratings, driven primarily by local colour and texture information rather than global shape.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.