Home

TANSAV: Temporal Analysis of Neural Structures for Artificial Vision

TENSAV Artificial Vision

TANSAV is designed and optimized primarily for visual object recognition. However, it's intended to be adaptable to other types of sensory input (e.g., audio, tactile). In addition to image inputs, it has the capability to input sets of numerical values.

The goal for TANSAV is to be able to emulate complex neuronal networks on a desktop PC or workstation with acceptable training times. The conflicting requirements of high-performance, large network sizes and minimal memory usage necessitates algorithms and data structures that are very complex.

TANSAV can model neurons ranging from simple point cells up to complex variable-state models. Networks can have multiple layers of sources and nodes with many connectivity options. As is typical, training the network is done by presenting repeated examples over multiple epochs using associative learning. The expectation is that the unique capabilities of the program will allow the number of training epochs to be fewer than is typical for standard ANN programs, using much less energy.

Temporal Analysis of Neural Structures for Artificial Vision

TANSAV is the name for a suite of three large software programs developed for machine learning vision applications.

The name describes the basic function of the program suite. However, TANSAV is not yet another "deep learning" or "spiking" simulator. The program uses learning, but is designed to have a combination of features and capabilities that are not available on any existing simulators.

The TANSAV package consists of the software code for the programs, a 300+ page illustrated User Manual , and a 1200+ page illustrated programmer's Design Document. .

Development and testing is ongoing, but the operation of the programs has been heavily tested and debugged sufficiently to declare them at "beta" stage.

The About page gives a limited description of TANSAV and the Background page describes the motivations for pursuing it.

Project Goals