Emergence of foveal image sampling from learning to attend in visual scenes
This work addresses how visual attention mechanisms can be optimized for tasks like object detection in cluttered scenes, though it is incremental in exploring learned sampling properties.
The paper tackled the problem of learning efficient visual attention by training a neural model with a learnable retinal sampling lattice on a visual search task, resulting in the emergence of a foveal-like sampling pattern that resembles the primate retina's eccentricity-dependent structure.
We describe a neural attention model with a learnable retinal sampling lattice. The model is trained on a visual search task requiring the classification of an object embedded in a visual scene amidst background distractors using the smallest number of fixations. We explore the tiling properties that emerge in the model's retinal sampling lattice after training. Specifically, we show that this lattice resembles the eccentricity dependent sampling lattice of the primate retina, with a high resolution region in the fovea surrounded by a low resolution periphery. Furthermore, we find conditions where these emergent properties are amplified or eliminated providing clues to their function.