Edge encoding mechanisms in the parallel L4 neuron array of the fly (Musca domestica).

The current model of fly's cartridge termed Olson's Algorithm employs seven photoreceptors (R1-R6, Rref) to encode edge orientation. The models provide a sinusoidal output for an edge rotated about the reference photoreceptor. Although the model is powerful in providing insight into a fly's visual apparatus, it has several shortcomings. The model: does not respond to horizontal edges, does not provide Gaussian overlapped weighting of the photoreceptors which provides hyperacuity capability, and does not provide for a unique encoding of edge information. We are currently working on modifications to the model to remedy these issues. These issues must be resolved prior to pursuing object segmentation, tracking, and identification. In this paper we will investigate enhancements to the model and what they provide in solving some of the issues.