I3D 2011 Report: – Part I: Keynote, Part II: Industry Session. In Part II, Chris Hecker has a game developer’s wish list for researchers.

CVPR 2011 program is available. The acceptance rate of papers by area is listed at the bottom of this web page.

COMPUTATIONAL PHOTOGRAPHY AND VIDEO, at MIT.

Returning Control to the Programmer: SIMD Intrinsics for Virtual Machines, in ACM Queue Feb 2011.

The Multicore Future, to be a five-part series on the research in MIT. Part of the research is done within the Angstrom project.

A Cost-Efficient L1–L2 Multicore Interconnect: Performance, Power, and Area Considerations, from IBM, in TCSI March 2011. A lot of effort is spent on analysing the activity factor of the components for optimal choice of threshold voltage.

On the Roots of Wireless Communications, in Circuits and Systems Magazine 2011. The story starts in 1800s and ends in the 1920s.

Two interesting papers in Computer Graphics and Applications March-April 2011:

• Colorization Using the Rotation-Invariant Feature Space, rotation-invariant Gabor filter and feature space optimization is used.
• A Virtual Reality Simulator for Ultrasound-Guided Biopsy Training, stitching volumetric data using 3D SIFT.

Fast Communication Architecture Exploration of Processor Pool-Based MPSoC via Static Performance Analysis, in TCAD March 2011. Its trace-based bandwidth analysis and memory contention analysis seem to be accurate for random traffic. It is unknown whether it will be accurate for real application. The architecture template in the paper has coarse grained contention. It is unknown whether the contention analysis will work for multiple banked SRAMs in an SIMD cluster, e.g. in GPUs.

Two-Degree-of-Freedom Active Vibration Control of a Prototyped “Smart” Rotor, from Delft, in IEEE Transactions on Control Systems Technology March 2011. The design methodology is well discussed in the paper. Both first principles modelling (physical modelling) and experimental modelling (system identification) are used.

Coded Strobing Photography: Compressive Sensing of High Speed Periodic Videos, in TPAMI April 2011. It works for periodic video sequences. However, it does have some pointers to super-resolution techniques. For example, it points to two papers on the fundamental limitations of super-resolution:

• Limits on super-resolution and how to break them, in TPAMI 2002.
• Fundamental limits of reconstruction-based superresolution algorithms under local translation, in TPAMI 2004.
• Comments on “Fundamental limits of reconstruction-based superresolution algorithms under local translation”, in TPAMI 2006.

Some other references on super-resolution:

• Motion deblurring and super-resolution from an image sequence , in ECCV 96.
• Example-Based Super-Resolution, in Computer Graphics and Applications 2002.
• MRI inter-slice reconstruction using super-resolution, in MRI 2002.
• Improving resolution by image registration, in 1994.

There are other interesting paper in TPAMI March 2011:

• Context-Dependent Kernels for Object Classification. SVM with context-aware kernels has a clear win over SVM with context-free kernels.
• Dynamic Programming and Graph Algorithms in Computer Vision. These algorithmic patterns should be supported in a vision processors.
• Linearized Motion Estimation for Articulated Planes, articulated constrains are applied to gradient based motion estimation for an affine camera and feature based motion estimation for a projective camera.
• Robust Stereo Matching Using Adaptive Normalized Cross-Correlation
• Topology-Adaptive Mesh Deformation for Surface Evolution, Morphing, and Multiview Reconstruction

Signal Processing Magazine March 2011 has a special issue on Compressive Sensing. The Signal Processing Magazine also has some interesting articles on giving presentation:

• Effective Communication: Excellence in a Technical Presentation [DSP Education], in March 2008.
• Storytelling—The Missing Art in Engineering Presentations [DSP Education], in March 2011.

Two paper in IISWC 2009 on workload characterization of feature detection:

• Image feature extraction for mobile processors
• Performance characterization and optimization of mobile augmented reality on handheld platforms

Some interesting papers in TVLSI March 2011:

• High Performance and Area Efficient Flexible DSP Datapath Synthesis. It uses an architecture template with flexible pipeline stages (FPS) and both horizontal and vertical parallelism.
• On-Chip Interconnect Analysis of Performance and Energy Metrics Under Different Design Goals. However, usually the designers do not know the design goals of the interconnection. The designers only know the design goals of the whole system. For example, a complete system optimized for power may need an interconnection optimized for delay.

Compressive Rendering: A Rendering Application of Compressed Sensing, in TVCG April 2011. It reduces the number of pixels to be processed by 30% without compromising too much quality.

Designing Network On-Chip Architectures in the Nanoscale Era, a recent book.

A Refactoring Approach to Parallelism, in IEEE Software Jan-Feb 2011. The authors believe, while automatic parallelization may work for small kernels, interactive parallelization is required for large-scale, irregular, non-scientific software.

Robust visual-servo control of robot manipulators in the presence of uncertainty, an old paper in CDC 1999. The robustness under uncertainty is well analysed in this paper. This paper is widely cited over the years. The latest citations are from two papers in ACC 2010:

• Image-based Robust Control of Robot Manipulators using Dynamic Compensator
• Visual closed-loop tracking with area stabilization

The ParLab called their many-core OS the “ROS” in the paper “Processes and Resource Management in a Scalable Many-core OS“. It may confuse people who knows another ROS, which is short for “Robot Operating System”. The RAMP and FAME project have some publications in 2010:

• RAMP gold: an FPGA-based architecture simulator for multiprocessors, in DAC 2010.
• A case for FAME: FPGA architecture model execution, in ISCA 2010.

Scalable instruction set simulator for thousand-core architectures running on GPGPUs, in HPCS 2010. It performs ISA simulation, with GPU as host and CMP as target.

How to simulate 1000 cores, in SIGARCH News March 2009. It performs trace simulation, with CMP as host and CMP as target.

Some related works of “How to simulate 1000 cores” from google scholar.

It is funny that, when I search “concurrent data structure” in google scholar, the first result is the PhD thesis of Williams J. Dally, titled “A VLSI Architecture for Concurrent Data Structures“, which was written in 1986. Classics are timeless.

Some interesting articles in Communications of ACM March 2010:

• Computer and information science and engineering: one discipline, many specialties. I did not notice the name of the author until I finished the whole article. I thought it was written by someone in fields like philosophy of science. It turns out to be written by Marc Snir. He says the term “mathematical engineering” comes from the Dutch language. I guess he means the Dutch term “wiskundige ingenieurstechniek“. Mathematical engineering “creates mathematics under constrains”. For computer engineering, the obvious constrain is the physical constrain, under which “mathematical abstractions interplay with their physical embodiment”. Marc, however, believes “physical constrains are a small fraction of the constrains relevant to the design of computing and information systems”.
• Data structures in the multicore age, by Nir Shavit. He has a book chapter on concurrent data structures in 2001, and a book titled “The Art of Multiprocessor Programming” in 2006.
• Do you own the software you buy? No, in short.
• Managing time. Lack of time is the symptom, not the problem.

Global Contrast based Salient Region Detection, to appear in CVPR2011.

Efficient Smart CMOS Camera Based on FPGAs Oriented to Embedded Image Processing, in Sensors. The system uses an Aptina mt9m413c36stm image sensor, which is packaged with 280-pin PGA. An off-the-shelf FPGA development board usually does not have so many free pins. The system is likely to use a customized board.

Journal of Robotics Research March 2011 has a Special Issue on the Fourteenth International Symposium on Robotics Research, 2009.

Journal of Physical Communication has a call for paper for “Special Issue on Cognitive Radio for LTE Advanced & Beyond”. The call for paper itself is a good survey on the state-of-the-art.

Off topic:
The problem of money as a measuring stick, in ACM crossroad Spring 2011. It tries to use money to solve the problem of money. According to Einstein, “The significant problems we have cannot be solved at the same level of thinking with which we created them”. Money is often deliberately manipulated for economic and political reasons, which makes it the worst candidate as a measuring stick. There should be better measuring sticks than the monetary system.

The proceeding of Conference on Decision and Control (CDC) 2010 is released. An overview of the papers is available here..

The Hendrik W. Bode Lecture: The role of theory in control practice, by Manfred Morari, in CDC 2010. To obtain the slides, a form need to be filled in.

CVPR 2011 papers on the web, a web page for early birds.

Processing, a programming language and environment for visualization. There are short tutorials in the IBM developerWorks: “Data visualization with Processing, Part 1, Part 2, Part 3“.

Off topic:
Information overload? Time to relax then. The author believes important information is likely to be redundant and artificially selected. However, reproduction and artificial selection of information leads to slow convergence. For people who need to quickly react to the dynamics of the information, that may not be an option.

What is a Computational Camera?, a web page created in the computer vision lab of Columbia University. It has a recent survey “Computational Cameras: Approaches, Benefits and Limits” published in Jan 2011.

Dynamic and Robust Streaming in and between Connected Consumer-Electronic Devices, a book published in 2005 by the people from Philips.

Eclipse: A Heterogeneous Multiprocessor Architecture for Flexible Media Processing, PhD thesis of Martijn Rutten.

C-HEAP: A Heterogeneous Multi-Processor Architecture Template and Scalable and Flexible Protocol for the Design of Embedded Signal Processing Systems , in Design Automation for Embedded Systems 2002.

Journal of Imaging Systems and Technology March 2011 has a special issue on Combinatorial problems and algorithms in image analysis.

Large Displacement Optical Flow: Descriptor Matching in Variational Motion Estimation, TPAMI March 2011.

SIFT Flow: Dense Correspondence across Scenes and its Applications, matching based on dense SIFT.