Unlike existing global shape-from-shading (SFS) algorithms which involve the brightness constraint in their formulation, we propose a new SFS algorithm which replaces the brightness constraint with an intensity gradient constraint. This is a global approach which obtains the solution by the minimization of an error function over the entire image. Through the linearization of the gradient of the...

With the fast development of deep learning, people have started to train very big neural networks using massive data. Asynchronous Stochastic Gradient Descent (ASGD) is widely used to fulfill this task, which, however, is known to suffer from the problem of delayed gradient. That is, when a local worker adds the gradient it calculates to the global model, the global model may have been updated ...

Recurrent neural networks (RNNs), with the capability of dealing with spatio-temporal relationship, are more complex than feed-forward neural networks. Training of RNNs by gradient descent methods becomes more dii-cult. Therefore, another training method, which uses cellular genetic algorithms, is proposed. In this paper, the performance of training by a gradient descent method is compared with...

We combine in this paper the topological gradient, which is a powerful method for edge detection in image processing, and a variant of the minimal path method in order to find connected contours. The topological gradient provides a more global analysis of the image than the standard gradient and identifies the main edges of an image. Several image processing problems (e.g., inpainting and segme...

We combine in this paper the topological gradient, which is a powerful method for edge detection in image processing, and a variant of the minimal path method in order to find connected contours. The topological gradient provides a more global analysis of the image than the standard gradient, and identifies the main edges of an image. Several image processing problems (e.g. inpainting and segme...

We propose an ultra low cost 2D-to-3D conversion system which generates depth maps using global scene depth gradient and texture-based local depth refinement. The depth gradient strength of global scene is generated by only analyzing the edge feature horizontally. Then the local pixel value, Y, Cb, and Cr component of the video content is used to refine the detail depth value. Combining the glo...

The paper is focussing on some recent developments in nonparametric mixture distributions. It discusses nonparametric maximum likelihood estimation of the mixing distribution and will emphasize gradient type results, especially in terms of global results and global convergence of algorithms such as vertex direction or vertex exchange method. However, the NPMLE (or the algorithms constructing it...

This paper presents a fast algorithm for global motion estimation based on Iterative LeastSquare Estimation (ILSE) technique. Compared with the traditional framework, three improvements were made to accelerate the computation progress. First, a new 3-parameter linear model, together with its solution using modified ILSE method, is proposed to describe and estimate global motion, which is simple...