Yuebing Jiang

Dynamically reconfigurable DCT architectures based on bitrate, power, and image quality considerations

We propose a dynamically reconfigurable DCT architecture system that can be used to optimize performance objectives while meeting real-time constraints on power, image quality, and bitrate. The proposed system can be dynamically reconfigured between 4 different modes: (i) minimum power mode, (ii) minimum bitrate mode, (iii) maximum image quality mode, and (iv) typical mode. The proposed system relies on the use of efficient DCT implementations that are parameterized by the word-length of the DCT transform coefficients and the use of different quantization quality factors. Optimal DCT architectures and quality factors are pre-computed on a training dataset. The proposed system is validated on the LIVE database using leave-one-out. From the results, it is clear that real-time constraints can be successfully met for the majority of the test images while optimizing for the 4 modes of operation.

A dynamically reconfigurable architecture system for time-varying image constraints (DRASTIC) for motion JPEG

We propose a dynamically reconfigurable system for time-varying image constraints (DRASTIC) for applications in video communications. DRASTIC defines a framework for both joint and independent optimization of dynamic power, image quality, and bitrate subject to different constraint scenarios. We demonstrate DRASTIC for intra-mode video encoding for MJPEG. However, since the DCT is critical component of most video coding standards, our approach can be extended to modern standards such as AVC (H.264), and emerging standards such as HEVC (H.265), and VP9. Based on a hardware–software co-design approach, we define a family of scalable 2D DCT hardware modules that are jointly optimized with the quality factor (in software). We generate a total of 1,280 configurations of which 841 were found to be Pareto optimal. For full 2D DCT calculation, the results indicate that the proposed method is DRASTIC mode implementation at least as good or significantly better than any previously published implementation. A scalable, real-time controller is used for selecting an appropriate configuration so as to meet time-varying constraints. The real-time controller is shown to satisfy the constraints of different communications modes (e.g., minimum dynamic power, maximum image quality, etc.) as well as to adapt to mode changes. Empirically, we have found that the DRASTIC controller adapts to meet the new constraints within five video frames of a mode change. Overall, the proposed approach yields significant savings over the use of comparable static architectures.

Cell-Based Architecture for Adaptive Wiring Panels: A First Prototype

Wepresent a first prototype for developing the concept of amanifold of adaptive wiring cells connected as a single overall adaptive wiring panel. Themain use of the adaptive wiring panel is related to affordable plug-and-play space applications. A reconfigurable switch fabric enables dynamic routing of signals and power; thus, power, digital, and analog signals can be routed for space systems. This concept can also be applied to terrestrial applications such as aircraft wiring and ground-based systems. The adaptive wiring panel is a manifold of adaptive wiring cells cast as a single overall panel. The panel is a pegboard-like structure, which does not articulate specific sockets, but rather provides a continuous grid of contact pads and mechanical mounting holes. Implementation is based on three basic elements: 1) cell units, which are minimum independent units of the adaptive wiring panel, each with interconnections and links with other cells to form the switch fabric by which wewire components to each other; 2) a cell management unit, which talks independently with all cell units and manages the wiring path and panel switch connections; and 3) modules that provide the components to be wired.

JPEG image compression using quantization table optimization based on perceptual image quality assessment

We consider the use of perceptual image quality assessment for quantization table (QT) optimization for JPEG compression. For evaluating performance, we consider the use of the Structural Similarity Index (SSIM) for evaluating distortion in the compressed images. This leads to the study of rate-SSIM curves that replace the traditional use of rate-distortion curves based on the PSNR.

A unified and pipelined hardware architecture for implementing intra prediction in HEVC

The High Efficiency Video Coding (HEVC) standard can achieve significant improvements in coding performance over H.264/AVC. To achieve significant coding improvements in intra-predictive coding, HEVC relies on the use of an extended set of intra-prediction modes and prediction block sizes. This paper presents a unified hardware architecture for implementing all 35 intra-prediction modes that include the planar mode, the DC mode, and all angular modes for all prediction unit (PU) sizes ranging from 4 × 4 to 64 × 64 pixels. We propose the use of a unified reference sample indexing scheme that avoids the need for sample re-arrangement suggested in the HEVC reference design. The hardware architecture is implemented on a Xilinx Virtex 5 device (XC5VLX110T) for which we report power measurements, resource utilization, and the average number of required cycles per pixel.

Scalable Open-Source Architecture for Real-Time Monitoring of Adaptive Wiring Panels

The first prototype of an adaptive wiring panel was recently introduced that implemented a reconfigurable switch fabric that allows dynamic routing of analog, digital, and power signals for space system applications. In this paper, a complete redesign and reimplementation of the adaptive wiring panel system is considered to address issues associated with scalability, reliability, and real-time monitoring of the switching fabric. The new system is demonstrated using 48 cells as opposed to the six cells of the first adaptive wiring panel prototype. The hardware and software systems are open source, and recommendations are provided to support further extensions to the system.

Dynamically reconfigurable architecture system for time-varying image constraints (DRASTIC) for HEVC intra encoding

We introduce the use of a dynamically reconfigurable architecture system for time-varying image constraints (DRASTIC) and consider applications in HEVC intra encoding. DRASTIC provides a framework for jointly optimizing complexity-rate-distortion for different operating modes. DRASTIC optimization involves dynamically reconfiguring parameters (e.g., the quantization parameter, encoding configuration modes) in software for achieving fine optimization control. The fine control achieved with the use of the DRASTIC modes is shown to perform significantly better than the standard use of fixed profiles.

A dynamically reconfigurable DCT architecture for maximum image quality subject to dynamic power and bitrate constraints

We introduce a dynamically reconfigurable DCT architecture system that allows us to select optimal implementations based on bitrate and dynamic power constraints. For our approach, we first compute Pareto-optimal hardware realizations that are assessed in terms of reconstructed image quality, bitrate and dynamic power. The space of Paretooptimal realizations are generated by varying both the number of non-zero DCT coefficients and the quality factor for the quantization table. From the generated hardware realizations, we then select the Pareto-optimal cases and discard all other cases. For each bitrate and dynamic power constraint, we use a dynamic partial reconfiguration (DPR) controller to implement the optimal DCT architecture. We test our approach using leave-one-out on the LIVE database, and implement our system on a Virtex-5 FPGA and demonstrate its performance using different bitrate and dynamic power constraints.

Scalable HEVC intra frame complexity control subject to quality and bitrate constraints

We introduce an optimal approach for minimizing the computational complexity of HEVC intra encoding subject to constraints in bitrate and reconstruction quality. Our constraint-optimization approach provides an extension to the use of bit constrained rate-distortion optimization so as to minimize encoding time while also delivering sufficient video quality. For our approach, we adaptively control the quantization parameter (QP) and a quad-tree-depth oriented coding tree unit configuration to deliver performance that is optimal in the complexity-rate-quality performance space. Our proposed approach employs a spatially adaptive model that uses neighboring configurations to estimate optimal values for QP and the coding tree unit configuration. We tested our approach using an HEVC standard test video and the ability to dynamically reconfigure between low, medium and high profiles. We found that we can meet the constraints (at 93% (low), 83% (medium) and 93% (high)), while delivering encoding time savings of 13%, 49% and 40% respectively.

A configurable system for role-specific video imaging during laparoscopic surgery

In minimally invasive surgery, the surgeon and the assistant rely on a single laparoscopic video view for performing different clinical roles. The assistant is tasked with manipulating the camera view so as to maintain a global, panoramic view of the operation. The surgeon needs to remain focused on the operation, requiring a detailed close-up view. We use the term role-specific video imaging to describe the need to provide separate views for the assistant and the surgeon. In this paper we introduce role-specific video imaging for laparosopic surgery. The system is designed to be configurable in the sense that imaging parameters and algorithms can be adjusted in real-time so as to meet the specific needs that arise. The system was evaluated on 4 cases by two surgeons on a Linux-based 3.2.0 Kernel, with 4GB RAM, and Intel 3.4GHz I7 (2nd generation) microprocessor. Clinical evaluation of the different configuration modes has helped us determine that high-quality role-specific imaging can be achieved for zooming factors that are larger or equal to 2×2 with bilinear interpolation, while maintaining 30 frame per seconds for the panoramic and close-up views. In future work, in order to minimize interaction with the surgical team, the system will be upgraded to incorporate tracking of the operating instrument during surgery.