Artificial neural network algorithms for some computer vision problems

The problem considered here involves the use of an artificial neural network for solving some computer vision problems such as static and motion stereo, computation of optical flow, and image restoration. The network used in this research contains massive mutually interconnected and self-connected binary neurons. Two decision rules, deterministic and stochastic decision rules, are designed to ensure convergence. The stochastic decision rule guarantees convergence to a global minimum but computationally is very expensive. While the deterministic decision rule greatly reduces solution time, gives only a local minimum. Two basic methods, static and motion stereo, are considered. The intensity derivatives which are more reliable than conventional measurement primitives are used for matching. A window operator which functions very similar to the human eye in detecting the intensity changes is proposed for estimating the derivatives. With the epipolar, photometric and smoothness constraints the neural network is employed to implement the matching procedure. Batch and recursive algorithms which allow the use of arbitrarily many image frames are presented for motion stereo. No surface interpolation step is involved in all algorithms because of the dense derivatives used. An algorithm for computing optical flow using rotation invariant primitives extracted from successive monocular images is presented. Under the local rigidity assumption and the smoothness constraint, the neural network is used to compute optical flow. To locate motion discontinuities, the information about occluding elements are utilized by embedding it into the bias inputs of the network. A batch solution is also developed for the case of pure translation. An algorithm for restoration of gray level images degraded by a known shift invariant blur function and additive noise is developed. The neural network is employed to represent a possibly nonstationary image whose gray level function is the simple sum of the neuron state variables. The nonlinear restoration method is carried out iteratively by using a dynamic algorithm to minimize the energy function of the network. Owing to the model's fault-tolerant nature and computation capability, a high quality image is obtained using this approach. A practical algorithm with reduced computational complexity is also presented. The choice of the boundary values to reduce the ringing effect is discussed and comparisons to other restoration methods are given. A schematic diagram of optical implementation of the restoration algorithm is described. (Copies available exclusively from Micrographics Department, Doheny Library, USC, Los Angeles, CA 90089-0182.)