We consider convex optimization problems with structures that are suitable for sequential treatment or online sampling. In particular, we focus on problems where the objective function is an expected value, and the constraint set is the intersection of a large number of simpler sets. We propose an algorithmic framework for stochastic first-order methods using random projection/proximal updates ...

We propose to solve an image segmentation problem with connectivity constraints via projection onto the constraint set. The constraints form a convex set and the convex image segmentation problem with a total variation regularizer can be solved to global optimality in a primal-dual framework. Efficiency is achieved by directly computing the update of the primal variable via a projection onto th...

The cardinality constraint is an intrinsic way to restrict the solution structure in many domains, for example, sparse learning, feature selection, and compressed sensing. To solve a cardinality constrained problem, the key challenge is to solve the projection onto the cardinality constraint set, which is NP-hard in general when there exist multiple overlapped cardiaiality constraints. In this ...

In continuous wave (CW) radar and high pulse repetition frequency pulse-Doppler (HPRF-PD) radar, the interference plus noise sample snapshots are hard to be obtained. The desired signal in the received snapshots makes the LCMV-based adaptive monopulse algorithm sensitive to pattern look direction error. A linearly constrained subarray robust adaptive monopulse algorithm based on main lobe maint...

This paper presents a general method for deriving approximate dynamical equations that satisfy a prescribed constraint typically chosen to filter out unwanted highfrequency motions. The approximate equations take a simple general form in arbitrary phase-space coordinates. The family of semigeostrophic equations for rapidly rotating flow derived by Salmon (1983, 1985) fits this general form when...

This paper presents a new iterative equalization algorithm that maximizes capacity for discrete multitone (DMT) systems. The research modifies a previously proposed criterion and applies an appropriate transformation to map the objective function and the constraint set into a canonical region. The resulting constraint set exhibits an identifiable geometric characteristic. Using the gradient pro...

We propose an edge-preserving filtering method with a novel use of the L0 gradient. Our method, termed as the L0 gradient projection, is formulated as the minimization of a quadratic data-fidelity to an input image subject to the constraint that theL0 gradient, the number of non-zero gradients, of the output image is less than a user-given parameter α. This strategy is much more intuitive than ...

Multi-label classification has attracted an increasing amount of attention in recent years. To this end, many algorithms have been developed to classify multi-label data in an effective manner. However, they usually do not consider the pairwise relations indicated by sample labels, which actually play important roles in multi-label classification. Inspired by this, we naturally extend the tradi...

Sequential pattern mining under constraints is a challenging data mining task. Many efficient ad hoc methods have been developed for mining sequential patterns, but they are all suffering from a lack of genericity. Recent works have investigated Constraint Programming (CP) methods, but they are not still effective because of their encoding. In this paper, we propose a global constraint based on...

Fisheye lenses are often used to enlarge the field of view (FOV) of a conventional camera. But the images taken with fisheye lenses have severe distortions. This paper proposes a novel calibration method for fisheye lenses using images of space lines in a single fisheye image. Since some fisheye cameras’ FOV are around 180 degrees, the spherical perspective projection model is employed. It is w...