Seminar Series

Low level visual processing and face recognition

I consider the idea that the extraordinary range of useful information we extract from human faces means that faces are optimized for human visual processing. First, I will present evidence that facial identity information is conveyed largely via mechanisms tuned to horizontal visual structure (suggesting faces could be optimized for processing in the earliest visual cortical areas). Specifically observers are better at identifying faces that contain only horizontal information, compared to any other orientation band. I will also show that horizontal structures within faces have an unusual tendency to fall into vertically co-aligned clusters (“bar-codes”). I will propose that it is this property makes faces “special” visual stimuli because they are able to transmit information within a constrained one-dimensional code. This structure affords computational advantages for face detection and decoding, including robustness to normal environmental image degradation, but makes faces vulnerable to certain classes of transformation that change the sequence of bars such as spatial inversion or contrast-polarity reversal. I will also consider face adaptation, and in particular show that these phenomena are tightly tuned to basic visual aspects of the stimulus such as its size, scale and spatial frequency content