A Computational Approach to Edge Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Contour Detection and Hierarchical Image Segmentation
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Mathematical Model of Retinal Ganglion Cells and Its Applications in Image Representation
Neural Processing Letters
Hi-index | 0.00 |
Orientation detection is an important step in image understanding. Simple cells in the primary visual cortex are sensitive to stimuli of specific orientation. Hubel and Wiesel's classical theory attributes orientation selection in a simple cell to the aligned receptive fields of the retinal ganglion cells and the lateral geniculate nucleus cells. It is hypothesized that these receptive fields are similar in size and sub-structure. Current anatomical and electrophysiological evidence is not sufficient to support this geometric-arrangement-based model. Additionally, this feedforward model is incomplete in the implementation details of neural computation. The proposed algebraic model satisfies local constraints from afferent neurons and estimates the orientation of the stimulus. This constraint satisfaction based approach achieves desirable results on challenging image dataset, and can also be used to explain visual illusions.