Sorry, you need to enable JavaScript to visit this website.

In this paper, we propose a video haze removal system that can be used to assist vehicle drivers and autonomous vehicles in hazy weather conditions. Our design objectives are to provide a high speed system with high perceived resolution that can be integrated in vehicle dashboard. We achieved these objectives by targeting TMS320DM6446 platform, where we propose a distribution of the tasks of the algorithm among the heterogeneous processor cores: ARM, DSP, and VICP.

A fast and high-quality depth estimation algorithm was proposed. Proposed algorithm accelerated depth estimation speed by 68-116 times faster than a Graph-Cuts algorithm. Estimated depth map quality is +0.3 to -1.0dB higher or lower than Graph-Cuts algorithm for super multi-view image synthesis. Depth estimation time is determined by the image size and number of depth layers, which is useful for implementation.

Inverse problems encountered in video processing often require to minimize criteria involving a high number of variables. Among available optimization techniques, proximal methods have shown their efficiency in solving large-scale possibly nonsmooth problems. When some of the proximity operators involved in these methods do not have closed form expressions, they may constitute a bottleneck in terms of computational complexity and memory

Accurate and robust spot tracking is a necessary tool for quantitative motion analysis in fluorescence microscopy images. In this work, we exploits the underlying stationary motion in biological systems, e.g. the movement of crowds, bacteria swarming and cyclosis in plant cells, and then propose a multi-frame optical flow based tracker. We obtain the stationary motion by adapting a recent optical flow algorithm that relates one image to another locally using an all-pass filter.

In this work, a 3D modelization of the surrounding environment is enabled with an improvised ad-hoc camera networks of both static and mobile devices (cloud vision network).

A gray-level image can be represented by a component tree, based on the inclusion relation of connected regions obtained by threshold decomposition. The great advantage of this structure is the efficient determination of a set of attributes for each component of the image, being widely used in morphological filtering (for instance, area as attribute to the area opening).