Achieving robust face recognition from video by combining a weak photometric model and a learnt generic face invariant

In spite of over two decades of intense research, illumination and pose invariance remain prohibitively challenging aspects of face recognition for most practical applications. The objective of this work is to recognize faces using video sequences both for training and recognition input, in a realistic, unconstrained setup in which lighting, pose and user motion pattern have a wide variability and face images are of low resolution. The central contribution is an illumination invariant, which we show to be suitable for recognition from video of loosely constrained head motion. In particular there are three contributions: (i) we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation to exploit the proposed invariant and generalize in the presence of extreme illumination changes; (ii) we introduce a video sequence re-illumination algorithm to achieve fine alignment of two video sequences; and (iii) we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve robustness to unseen head poses. We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 323 individuals and 1474 video sequences with extreme illumination, pose and head motion variation. Our system consistently achieved a nearly perfect recognition rate (over 99.7% on all four databases). © 2012 Elsevier Ltd All rights reserved.

Patterning method to create a mask

A method of creating micro patterned devices by patterning thin films which are deposited on a substrate. A channel or channels is created on the substrate, the width being of fine enough resolution such that a flowable mask material can be drawn along the channel by capillary forces

A computational fluid-structure interaction study of inflation cycles of a tension cone decelerator

Inflatable aerodynamic decelerators present potential advantages for planetary entry in missions of robotic and human exploration. The design of these structures face many engineering challenges, including complex deformable geometries, anisotropic material response, and coupled shockturbulence interactions. In this paper, we describe a comprehensive computational fluid-structure interaction study of an inflation cycle of a tension cone decelerator in supersonic flow and compare the simulations with earlier published experimental results. The aeroshell design and flow conditions closely match recent experiments conducted at Mach 2.5. The structural model is a 16-sided polygonal tension cone with seams between each segment. The computational model utilizes adaptive mesh refinement, large-eddy simulation, and shell mechanics with self-contact modeling to represent the flow and structure interaction. This study focuses on the dynamics of the structure as the inflation pressure varies gradually, and the behavior of forces experienced by the flexible and rigid (the payload capsule) structures. © 2011 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

Behavior learning of a face robot based on the characteristics of human instruction

Computer simulation experiments were performed to examine the effectiveness of OR- and comparative-reinforcement learning algorithms. In the simulation, human rewards were given as +1 and -1. Two models of human instruction that determine which reward is to be given in every step of a human instruction were used. Results show that human instruction may have a possibility of including both model-A and model-B characteristics, and it can be expected that the comparative-reinforcement learning algorithm is more effective for learning by human instructions.

Fabrication of high quality two-dimensional photonic crystal mask layer patterns

This work discusses the fabrication of two-dimensional photonic crystal mask layer patterns. Photonic crystal patterns having holes with smooth and straight sidewalls are achieved by optimizing electron beam exposure doses during electron beam lithography process. Thereafter, to precisely transfer the patterns from the beam resist to the SiO2 mask layer, we developed a pulse-time etching method and optimize various reaction ion etching conditions. Results show that we can obtain high quality two-dimensional photonic crystal mask layer patterns.

A novel face recognition method based on hypersausage neural networks

In this paper, we firstly give the nature of 'hypersausages', study its structure and training of the network, then discuss the nature of it by way of experimenting with ORL face database, and finally, verify its unsurpassable advantages compared with other means.

AlGaInAs narrow stripe selective growth on substrates patterned with different mask designs

We have performed a narrow stripe selective growth of oxide-free A1GaInAs waveguides on InP substrates patterned with pairs of SiO2 mask stripes under optimized growth conditions. The mask stripe width varied from 0 to 40 mu m, while the window region width between a pair of mask stripes was fixed at 1.5, 2.5 or 3.5 mu m. Flat and smooth A1GaInAs waveguides covered by specific InP layers are successfully grown on substrates patterned with different mask designs. The thickness enhancement ratio and the photoluminescence (PL) spectrum of the A1GaInAs narrow stripe waveguides are strongly dependent on the mask stripe width and the window region width. In particular, a large PL wavelength shift of 79 nm and a PL FWHM of less than 64 meV are obtained simultaneously with a small mask stripe width varying from 0 to 40 mu m when the window region width is 1.5 mu m. We present some possible interpretations of the experimental observations in considering both the migration effect from a masked region and the lateral vapour diffusion effect.

Homoepitaxial growth and device characteristics of SiC on Si-face (0001) 6H-SiC

Homoepitaxial growth of SiC on a Si-face (0 0 0 1) GH-SIC substrate has been performed in a modified gas-source molecular beam epitaxy system with Si2H6 and C2H4 at temperatures ranging 1000 1450 degreesC while keeping a constant SiC ratio (0.7) in the gas phase. X-ray diffraction patterns, Raman scattering measurements. and low-temperature photoluminescence spectra showed single-crystalline SiC. Mesa-type SiC p-n junctions were obtained on these epitaxial layers, and their I-V characteristics are presented. (C) 2001 Elsevier Science B.V. All rights reserved.

Research of Face Detection System Applied for Detecting Faces with Complex Background and Illumination

An effective face detection system used for detecting multi pose frontal face in gray images is presented. Image preprocessing approaches are applied to reduce the influence of the complex illumination. Eye-analog pairing and improved multiple related template matching are used to glancing and accurate face detecting, respectively. To shorten the time cost of detecting process, we employ prejudge rules in checking candidate image segments before template matching. Test by our own face database with complicated illumination and background, the system has high calculation speed and illumination independency, and obtains good experimental results.

Research of Face Location System Based on Human Vision Simulations

In this paper we present a robust face location system based on human vision simulations to automatically locate faces in color static images. Our method is divided into four stages. In the first stage we use a gauss low-pass filter to remove the fine information of images, which is useless in the initial stage of human vision. During the second and the third stages, our technique approximately detects the image regions, which may contain faces. During the fourth stage, the existence of faces in the selected regions is verified. Having combined the advantages of Bottom-Up Feature Based Methods and Appearance-Based Methods, our algorithm performs well in various images, including those with highly complex backgrounds.

A face detection algorithm based on high dimensional space geometry

In this paper, a face detection algorithm which is based on high dimensional space geometry has been proposed. Then after the simulation experiment of Euclidean Distance and the introduced algorithm, it was theoretically analyzed and discussed that the proposed algorithm has apparently advantage over the Euclidean Distance. Furthermore, in our experiments in color images, the proposed algorithm even gives more surprises.

Face recognition based on PCA and multi-degree of freedom neurons

An algorithm of PCA face recognition based on Multi-degree of Freedom Neurons theory is proposed, which based on the sample sets' topological character in the feature space which is different from "classification". Compare with the traditional PCA+NN algorithm, experiments prove its efficiency.