Abstract
Object recognition by natural and artificial visual systems benefits from the identification of object boundaries. A useful cue for the detection of object boundaries is the superposition of luminance and color edges. To gain insight into the suitability of this cue for object recognition, we examined convolutional neural network (CNNs) models that had been trained to recognize objects in natural images. Because CNNs are only trained to do a single task, any properties they possess are likely useful for that task. We focused specifically on units in the second convolutional layer invariant to contrast polarity, a useful trait for object boundary detection. Some of these units were tuned for a nonlinear combination of color and luminance, which is broadly consistent with a role in object boundary detection. Others were tuned for luminance alone, but few were tuned for color alone. A literature review reveals that V1 complex cells have a similar distribution of tuning. We speculate that this pattern of sensitivity provides an efficient basis for object recognition, perhaps by mitigating the effects of lighting on luminance contrast polarity. The paucity of contrast polarity-invariant representation of chromaticity alone suggests that it is redundant with other representations.
Funding: Funding: This work was supported by EY018849 grants to Gregory D Horwitz and EY07031 to Luke M Bun