Irregularization accelerates iterative regularization

When iterative methods are employed as regularizers of inverse problems, a main issue is the trade-off between smoothing effects and computation time, related to the convergence rate of iterations. Very often, faster methods obtain less accuracy. A new acceleration strategy is presented here, inspired by a choice of penalty terms formerly proposed in 2012 by Huckle and Sedlacek in the context of Tikhonov regularization by direct solvers. More precisely, we consider a special penalty term endowed with high regularization capabilities, and we apply it by using the opposite sign, that is negative, to its regularization parameter. This unprecedented choice leads to an “irregularization” phenomenon, which speeds up the underlying basic iterative method. The speeding up effects of the negative valued penalty term can be controlled through a sequence of decreasing coefficients as the iterations proceed in order to prevent noise amplification, tuning the weight of the correction term which generates the anti-regularization behavior. Filter factor expansion and convergence are analyzed in the simplified context of linear inverse problems in Hilbert spaces, by considering modified Landweber iterations as a first case study.     

Biological Effects of 25 to 150 GHz Radiation After In Vitro Exposure of Human Fibroblasts: a Comparison of Experimental Results

In this paper, we present a comprehensive discussion of the results obtained after in vitro exposure of human fetal fibroblasts and human adult fibroblasts to pulsed radiation in a wide band between 100 and 150 GHz and to continuous wave radiation at 25 GHz. In order to assess potential effects of exposure, the genome integrity, cell cycle, cytological ultrastructure, and proteins expression were evaluated.     

A new mesh-growing algorithm for fast surface reconstruction

This paper presents a new high-performance method for triangular mesh generation based on a mesh-growing approach. Starting from a seed triangle, the algorithm grows the triangular mesh by selecting a new point based on the Gabriel 2—Simplex criterion. This criterion can be considered to be a good approximation of the 2D Delaunay if the point cloud is well-sampled and not too rough. The performance of the proposed method is compared with that of the Cocone family and that of Ball Pivoting as regards the tessellation rate and the quality of the surface being generated from some benchmark point clouds and artificially noised test cases. The results are analysed and critically discussed.     

Effect of Surface Texture on the Adhesion Performance of Laser Treated Ti6Al4V Alloy

The growing demand in lighter and safer structures generates the requirement of lighter joining strategies, particularly for lightweight metal alloys, composites, and also joining dissimilar materials together. Titanium alloys stand out as the conventional choice for materials for light weight structures. Adhesive bonding of titanium is an appealing route for joint design, also for the possibility of joining it with dissimilar materials. The realization of a strong joint depends not only on the joint design and type of adhesive, but also on the preparation of the adhering surface. Laser texturing presents advantages compared to common surface preparation processes in terms of eco-compatibility, energetic efficiency, ease of manufacturing, and repeatability. This work presents a preliminary investigation on laser texturing of Ti6Al4 V alloy with a pulsed fiber laser source with the aim to increase surface adhesion for bonding. Particularly, different surface textures are proposed, and laser machining strategies are developed. The results showed that laser texturing provided up to eightfold and 30% higher shear strength compared to plain and sand blasted surfaces, respectively. Failure analysis showed that a margin of improvement is still possible by adapting the surface texture for better cavity filling and reducing surface damage caused by the laser treatment.     

Low-power DSP system for real-time correction of fish-eye cameras in automotive driver assistance applications

The development of an embedded system for real-time correction of fish-eye effect is presented. The fish-eye lens is applied to driver assistance video systems because of its wide-angled view. A large field of view can reduce the number of cameras needed for video system and their cost, installation, maintenance and wiring issues. On the other hand, this lens causes inherent radial distortion to image that has to be corrected in real-time with a low-cost and low-power processing platform. This paper proposes a solution that can be easily adapted to different types of lens and camera, and meets real-time constraints with a power budget within 100 mW and a board size of few cm². Starting from mathematical equations, given by the geometrical optics, a state-of-the-art correction method is presented, then optimizations are introduced at different levels: algorithmic level, where a real-time correction parameter calculation avoids extra non-volatile off-chip memory cards; data transfer level, where a new pixel pair management reduces memory access and storage burden; HW-SW implementation level, where a low-power board has been developed and tested in real automotive scenarios. Other applications of the developed system, such as multi-camera and multi-dimensional video systems, are finally presented.     

Modelling of back tempering in laser hardening

Back tempering is one of the most critical problems in laser hardening of extended surfaces. In this type of treatment, several laser tracks are slightly overlapped to obtain a uniform hardened surface. Due to the overlapping, tempered zones are generated on the treated surface with the consequent lack of uniformity in the surface hardness. In this work, a regression model was developed to estimate the loss of hardness due to the tempering effect as a function of the thermal cycle. A specific test, named laser surface treatment test, was designed and executed to reproduce the hardness reduction due to the tempering effect. An analytical thermal model was developed to evaluate the thermal cycle undergone by the material during this test. By the results of the laser surface treatment test combined with the analytical model, a prediction model was estimated. Good agreement was found between predicted and measured hardness decrease, and the identified model could be integrated in a numerical code to evaluate the optimal process parameters.