Temporally Dense Ray Tracing

We present a technique for real-time ray tracing with the goal of reaching 240 frames per second or more. The core idea is to trade spatial resolution for faster temporal updates in such a way that the display and human visual system aid in integrating high-quality images. We use a combination of frameless and interleaved rendering concepts together with ideas from temporal antialiasing algorithms and novel building blocks---the major one being adaptive selection of pixel orderings within tiles, which reduces spatiotemporal aliasing significantly.

A Fine-Grained GALS SoC with Pausible Adaptive Clocking in 16 nm FinFET

Modern SoCs suffer from power supply noise that can require significant additional timing margin, reducing performance and energy efficiency. Globally asynchronous, locally synchronous (GALS) systems can mitigate the impact of power supply noise, as well as simplify system design by removing the need for global timing closure. This work presents a 4mm2 distributed accelerator engine with 19 independent clock domains implemented in a 16nm process.

Voltage-Follower Coupling Quadrature Oscillator with Embedded Phase-Interpolator in 16nm FinFET

High-speed serial links require a very high frequency clock source. Multi-rings oscillator is a practical solution to this challenge. A new phase-interpolator embedded quadrature oscillator was designed and tested. Voltage-follower based cross-coupling loops create reliable and tunable phase relationship among OSC rings. The measurement results show that the proposed PI-OSC provides 1.25/0.97 LSB INL/DNL performance at 24GHz while consuming only 8.1mW power. This compact oscillator is suitable for clock generation in high-speed low-power links.

Wei Yang

Wei Yang is a Research Scientist of Robotics Research at NVIDIA, Seattle. His research interests include computer vision, machine learning, and their applications to robotics. He received his Doctoral degree in Electronic Engineering from the Chinese University of Hong Kong in 2018. Previously, he worked as a visiting scholar at Robotics Institute, Carnegie Mellon University (10/2017-4/2018). He received the masters degree in Computer Science and the BEng degree in Software Engineering from Sun Yat-sen University in 2014 and 2011, respectively. 

Yu-Wei Chao

Rowland O'Flaherty

Rowland's interest within robotics lie at the intersection of control theory, machine learning, and optimization. Rowland came to NVIDIA from the Silicon Valley follow-me drone startup world developing algorithms for planning and control, as well as improving the dynamical models of the robotic vehicles to a high degree of fidelity. Before joining the startup world, Rowland earned a B.S. and M.S degree in ECE from UCSB and a Ph.D. in Robotics from Georgia Tech.

Guillermo Marcus

Guillermo Marcus is a senior software engineer at NVIDIA at the intersection of graphics, communications, and machine learning. He is interested in systems engineering, software engineering and applied scientific computing with a focus on accelerated computing and application-specific accelerators. Before joining NVIDIA in 2013, he was a research scientist at the Univeristy of Heidelberg. He received the PRACE Award at ISC 2011. He has a Ph.D. (magna cum laude) from the Heidelberg University in Germany, a M.Sc.

Fabio Ramos

My research is focused on modelling and understanding uncertainty for prediction and decision making tasks, and includes Bayesian statistics, data fusion, anomaly detection, and reinforcement learning. Over the last ten years I have applied these techniques to robotics, mining and exploration, environment monitoring, and neuroscience.

Putting Humans in a Scene: Learning Affordance in 3D Indoor Environments

Affordance modeling plays an important role in visual understanding. In this paper, we aim to predict affordances of 3D indoor scenes, specifically what human poses are afforded by a given indoor environment, such as sitting on a chair or standing on the floor. In order to predict valid affordances and learn possible 3D human poses in indoor scenes, we need to understand the semantic and geometric structure of a scene as well as its potential interactions with a human. To learn such a model, a large-scale dataset of 3D indoor affordances is required.

SIDOD: A Synthetic Image Dataset for 3D Object Pose Recognition with Distractors

We present a new image dataset generated by the NVIDIA Deep Learning Data Synthesizer intended for use in object detection, pose estimation, and tracking applications. This dataset contains 144k stereo image pairs generated from 18 camera view points of three photorealistic virtual environments with up to 10 objects (chosen randomly from the 21 object models of the YCB dataset) and flying distractors. Object and camera pose, scene lighting, and quantity of objects and distractors were randomized. Each provided view includes RGB, depth, segmentation, and surface normal images.

Subscribe to