Brian Guenter

Microsoft · Microsoft Research

We present an approach to synthesize highly photorealistic images of 3D object models, which we use to train a convolutional neural network for detecting the objects in real images. The proposed approach has three key ingredients: (1) 3D object models are rendered in 3D models of complete scenes with realistic materials and lighting, (2) plausible...

The significant optical and size benefits of using a curved focal surface for imaging systems have been well studied yet never brought to market for lack of a high-quality, mass-producible, curved image sensor. In this work we demonstrate that commercial silicon CMOS image sensors can be thinned and formed into accurate, highly curved optical surfa...

We present a novel, automatic eye gaze tracking scheme inspired by smooth pursuit eye motion while playing mobile games or watching virtual reality contents. Our algorithm continuously calibrates an eye tracking system for a head mounted display. This eliminates the need for an explicit calibration step and automatically compensates for small movem...

Emerging uses of imaging technology for consumers cover a wide range of application areas from health to interaction techniques; however, typical cameras primarily transduce light from the visible spectrum into only three overlapping components of the spectrum: red, blue, and green. In contrast, hyperspectral imaging breaks down the electromagnetic...

Custom optics is a necessity for many imaging applications. Unfortunately, custom lens design is costly (thousands to tens of thousands of dollars), time consuming (10-12 weeks typical lead time), and requires specialized optics design expertise. By using only inexpensive, off-the-shelf lens components the Lens Factory automatic design system great...

A control system is described for controlling the operation of a target system, such as a data center. The control system uses a prediction module to predict demand for resources of the target system for future time steps. The control system then uses a transition determination module to determine state transitions within the target system to addre...

Humans see at high resolution only within a few degrees of the direction we are looking. Moving our gaze position around the current scene, we build up the rich visual landscape we consider reality from a sparse set of details, filled in with coarser data from our peripheral vision.

We exploit the falloff of acuity in the visual periphery to accelerate graphics computation by a factor of 5-6 on a desktop HD display (1920x1080). Our method tracks the user's gaze point and renders three image layers around it at progressively higher angular size but lower sampling rate. The three layers are then magnified to display resolution a...

All lenses have optical aberrations which reduce image sharpness. These aberrations can be reduced by deconvolving an image using the lens point spread function (PSF). However, fully measuring a PSF is laborious and prohibitive. Alternatively, one can simulate the PSF if the lens model is known. However, due to manufacturing tolerances lenses diffe...

We present ACES, an automated server provisioning system that aims to meet workload demand while minimizing energy consumption in data centers. To perform energy-aware server provisioning, ACES faces three key tradeoffs between cost, performance, and reliability: (1) maximizing energy savings vs. minimizing unmet load demand, (2) managing low power...

Most current linear-time forward dynamics algorithms support only simple types of joints due to difficulties in computing derivatives of joint transformations up to order two. We apply the D* symbolic differentiation algorithm to a recursive formulation of forward dynamics to get a highly efficient linear-time forward dynamics algorithm supporting...

Neon is a high-level domain-specific programming language for writing efficient image processing programs which can run on either the CPU or the GPU. End users write Neon programs in a C# programming environment. When the Neon program is executed, our optimizing code generator outputs human-readable source files for either the CPU or GPU. These sou...

Functions with densely interconnected expression graphs, which arise in computer graphics applications such as dynamics, space-time optimization, and PRT, can be difficult to efficiently differentiate using existing symbolic or automatic differentiation techniques. Our new algorithm, D*, computes efficient symbolic derivatives for these functions b...

This paper describes the application of space time constraints to creating transitions between segments of human body motion. The motion transition generation uses a combination of spacetime constraints and inverse kinematic constraints to generate seamless and dynamically plausible transitions between motion segments. We use a fast recursive dynam...

Skin is noticeably bumpy in character, which is clearly visible in close-up shots in a film or game. Methods that rely on simple texture-mapping of faces lack such high frequency shape detail, which makes them look non-realistic. More specifically, this detail is usually ignored in real-time applications, or is drawn in manually by an artist. In th...

This technical report details the complete dynamics formulation used in [3], including all the partial equations which could not appear in Siggraph due to space limitations. This tech-report is intended as a companion to that paper. 1 Equations of dynamics & their derivatives Constants, symbols, and notation:

Rendering realistic faces and facial expressions requires good mod- els for the reflectance of skin and the motion of the face. We describe a system for modeling, animating, and rendering a face using measured data for geometry, motion, and reflectance, which realistically reproduces the appearance of a par- ticular person's face and facial express...

High contrast images are common in night scenes and other scenes that include dark shadows and bright light sources. These scenes are difficult to display because their contrasts greatly exceed the range of most display devices for images. As a result, the image contrasts are compressed or truncated, obscuring subtle textures and details. Humans vi...

ing with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from Publications Dept, ACM Inc., fax + 1 (212) 869-0481, or <permissions@acm.org>. 1 Specializing Shaders Brian Guenter, Todd B. Knoblock, Erik Ruf * Microsoft Research...

We have created a system for capturing both the three-dimensional geometry and color and shading information for human facial ex- pressions. We use this data to reconstruct photorealistic, 3D ani- mations of the captured expressions. The system uses a large set of sampling points on the face to accurately track the three dimen- sional deformations...

This article presents a new lossless compression algorithm for computer animation image sequences. The algorithm uses transformation information available in the animation script and floating point depth and object number information at each pixel to perform highly accurate motion prediction with vary low computation. The geometric data (i.e., the...

: This paper introduces quadrature prefiltering, an accurate, efficient, and fairly simple algorithm for prefiltering polygons for scanline rendering. It renders very high quality images at reasonable cost, strongly suppressing aliasing artifacts. For equivalent RMS error, quadrature prefiltering is significantly faster than either uniform or jitte...

This paper presents a new lossless compression algorithm for computer animation image sequences. The algorithm uses transformation information available in the animation script and floating point depth and object number information stored at each pixel to perform highly accurate motion prediction with very low computation. The geometric data, i.e.,...

Kinematic simulation involves finding simultaneous zeros for a set of functions representing mechanical linkages which constrain the positions and orientations of a set of bodies. Evaluation of these functions, and of their Jacobians with respect to the positions and orientations of the bodies, is sufficient to allow the use of a multidimensional N...

g transmission bandwidth without significantly reducing image quality, or to improving the quality of signals compatible with the current NTSC standard. These bandwidth reduction and image improvement techniques are not directly applicable to the image synthesis problem, partially because some of them are concerned with overcoming the weaknesses of...

Stereopsis can provide essential depth information for recognizing objects and determining their spatial layout. Although some processing components are unique to stereopsis, lower monocular visual components can also affect stereopsis. For instance, we found an oblique effect for depth threshold, analogous to some monocular oblique effects. Are ex...

Specifying constraints on motion is simpler if the curve is parameterized by arc length, but many parametric curves of practical interest cannot be parameterized by arc length. An approximate numerical reparameterization technique that improves on a previous algorithm by using a different numerical integration procedure that recursively subdivides...

Specifying constraints on motion is simpler if the curve is parameterized by arc length, but many parametric curves of practical interest cannot be parameterized by arc length. An approximate numerical reparameterization technique that improves on a previous algorithm by using a different numerical integration procedure that recursively subdivides...

Computer simulation of human facial expression requires an interactive ability to create arbitrary face models and to control the simulated expreession on those models.The system described in this paper provides interactive facilities for attaching muscles and wrinkle lines to arbitrary rectangular face meshes and then for simulating the contractio...

Motion Compensated Noise Reduction Research Paper Regular Paper Stochastic sampling has proven to be an effective means of reducing many types of aliasing artifacts. However stochastic sampling requires many samples in order to produce images which appear free of noise, especially when used to antialias procedural textures or the shadowing effects...

Generative parametric CSG models, introduced by Snyder in 1992, have many desirable properties. Their functional representation is generally quite compact and resolution independent, since surfaces are represented as piecewise continuous functional programs. However, in practice they have proven impractical for real time rendering because of the di...

Generative parametric CSG models, introduced by Snyder in 1992, have many desirable properties. Their func- tional representation is generally quite compact and resolution independent, since surfaces are represented as piecewise continuous functional programs. However, in practice they have proven impractical for real time ren- dering because of th...

Derivatives arise frequently in graphics and scientific computation applications. As GPU's become more widely used for scientific computation the need for derivatives can be expected to increase. To meet this need we have added symbolic differentiation as a built in language feature in the HLSL shading language. The symbolic derivative is computed...

Functions with densely interconnected expression graphs, which arise in applications such as dynamics, space-time optimization, and PRT, can be difficult to efficiently differentiate using existing symbolic or automatic differentiation techniques. The new algorithm, D*, computes efficient symbolic derivatives for these functions by symbolically exe...

Directional filters applied to prefiltered imput image pyramids strike a good balance between fast computation and sharp, alias-free images, and are widely used for texture mapping and image warping. However, most published methods use implicit conversions between discrete and continuous images. This causes two flaws: 1) an input reconstruction fil...

Conventional mappings from the texture plane to three-dimensional surfaces fail to maintain the arc length of the texture when it is mapped onto the surface. This leads to unnatural texture appearance. In addition, texture gradient information is distorted, leading to poor perception of depth. We have developed an efficient algorithm which warps th...

It has long been known that the human visual system (HVS) is not a camera; the eye-retina-brain combination does not simply record a visual scene the way a movie camera or video camera does. Instead, the HVS processes the image in a series of steps of increasing abstraction. At each step some information about the scene may be lost and replaced by...

Does a pattern's orientation affect stereoacuity? Do vertical disparities play a role in stereopsis? Furthermore, is retinal disparity encoded by changes in spatial position or spatial phase? The latter implies a "stereo aperture problem" -- only disparities perpendicular to the pattern's orientation are encoded. We found stereoacuities are worse f...

The class of multidimensional signals we will be dealing with is three dimensional: x, y, and time. We want to compute the Fourier transform of a moving image and the easiest way to approach the problem is to begin by examining the behavior of affine transformations under the Fourier transform because image translation can be compactly described by...

Graphics, Visualization, and Usability (GVU) is an interdisciplinary area which draws its intellectual foundations from Computer Science, Psychology, Industrial and Systems Engineering, and Computer Engineering, and which has application to any use of computers to graphically convey information to users. Typical applications are computer aided desi...