Hyunchul Shin

An area-efficient VLSI architecture of a Reed-Solomon decoder/encoder for digital VCRs

A new flexible and area-efficient VLSI architecture of a Reed-Solomon product-code decoder/encoder has been developed for digital VCRs. The new architecture of the decoder/encoder targeted to reduce the circuit size and decoding latency has the following three features. First, a high area-efficiency has been achieved by sharing a functional block for encoding, modified syndrome computation, and erasure locator polynomial evaluation. Second, the circuit size and decoding latency has been reduced by using a new architecture to implement the modified Euclids algorithm. Third, by doubling the internal clock speed (from 18 MHz to 36 MHz), the decoding latency and hence the memory size can be reduced. The decoder/encoder designed by using the proposed method uses a reduced number of gates, by about 30%, than the one based on the conventional architectures.

A performance-driven logic emulation system: FPGA network design and performance-driven partitioning

FPGAs are widely used for logic emulation, software acceleration, custom computing, and prototyping. The architecture (or the interconnect mechanism of a FPGA network) of an emulator has profound effect on the performance (speed) and efficiency (chip utilization) of the emulator. In this paper, several architectures of FPGA networks are suggested, and they are compared with other typical existing architectures by using the MCNC partition benchmark circuits. Experimental results show that tripartite network outperforms six other typical architectures both in performance and in efficiency. For this study, the propagation delay of a path is estimated by the number of hops (interchip connections) and the number of intrachip connections on the path, and thus it is independent of a specific FPGA type. To partition a given circuit into the given prerouted network of FPGAs, a new routability-driven partitioning algorithm is developed. Experimental results using the MCNC benchmark examples show that our partition method produces better results than those of other recent approaches on the average, and that performance-driven partitioning is effective in reducing critical time delays.

Real-time single image dehazing using block-to-pixel interpolation and adaptive dark channel prior

The authors propose a novel and efficient method for single image dehazing. To accelerate the transmission estimation process, a block-to-pixel interpolation method is used for fine dark channel computation, in which the block-level dark channel is first computed, and then the fine pixel-level dark channel is obtained by a weighted voting of the block-level dark channel to preserve edges and smooth out texture noise. This technique can be used for a direct transmission map generation without a computationally expensive refinement step. Since the dark channel prior (DCP) is not valid in bright (sky) regions, they propose an adaptive DCP modelled by a Gaussian curve that produces a more natural recovered image of the sky and other bright regions. In addition, a scaling method for transmission map computation is proposed to further accelerate the dehazing method. Through experiments, they show that the proposed adaptive block-to-pixel technique is about 30 times faster and produces improved recovered images than the well-known state-of-the-art DCP approach.

Pedestrian detection and tracking using deformable part models and Kalman filtering

Both detection and tracking people are challenging problems, especially in complex real world scenes that commonly involve multi-person, complicated occlusions, and cluttered backgrounds. In this paper, we propose a novel approach for multi-person tracking-by-detection using deformable part models in Kalman filtering framework. The Kalman filter is used to keep track of each person and a unique label is assigned to each tracked individual. Based on this approach, people can enter and leave the scene at random. We test and demonstrate our results on the Caltech Pedestrian benchmark, which is the largest available dataset and consists of pedestrians varying widely in appearance, pose and scale. Complex situations such as people merging together are handled gracefully and individual persons can be tracked correctly after a group of people split. Experiments demonstrate the real-time performance and robustness of our system working in complex scenes. Our tracking model gives a tracking accuracy of 72.8% and a tracking precision of 82.3%.

Adaptive bilateral filtering for noise removal in depth upsampling

3D scene rendering requires depth maps and color information to produce high quality 3D results. Unfortunately, depth maps captured with the Time-of-flight (TOF) cameras have limited resolution and poor image quality, being severely influenced by the random and systematic noise, which makes them inapposite for generating high quality 3D images. In this paper, we have further analyzed a framework for upsampling the resolution of depth maps that jointly uses Gaussians of spatial and depth differences of low resolution depth maps pixels along with Gaussian of color intensity difference from high resolution 2D color image of the same scene. The variance of the Gaussian functions controls the amount of smoothing in uni-planner area and sharpness at boundaries. Using bigger variance smooths uni-planner area but blurs edges and vice versa. We have devised a method to adaptively calculate and use variance to get smoother surface and sharper edges of upsampled depth map with minimized noise.

New hand gesture recognition method for mouse operations

Hand gestures can be used for natural and intuitive human-computer interaction. Our new method combines existing techniques of skin color based ROI segmentation and Viola-Jones Haar-like feature based object detection, to optimize hand gesture recognition for mouse operation. A mouse operation has two parts, movement of cursor and clicking using the right or left mouse button. In this paper, color is used as a robust feature to first define a Region of Interest (ROI). Then within this ROI, hand postures are detected by using Haar-like features and AdaBoost learning algorithm. The AdaBoost learning algorithm significantly speeds up the performance and constructs an accurate cascaded classifier by combining a sequence of weak classifiers.

'Zone-refining' techniques for IC layout compaction

Zone-refining refers to a technique that forms a basis for layout compaction algorithms intermediate between one-dimensional (1-D) compactors and two-dimensional (2-D) placement techniques. An expanded zone in which 2-D refinement techniques are employed is repeatedly swept across the layout in different directions. The basic principle is reviewed and the computational complexity of zone refining is analyzed. The difficulties that had to be overcome in making the basic concept useful for compaction of integrated circuit layouts is discussed. One implementation is described, and some of the tradeoffs made and data structures used to obtain an efficient compactor are examined. The scope of possibilities for other implementations are discussed. >

Rendering high dynamic range images by using integrated global and local processing

We propose a new and effective tone reproduction technique that preserves the visibility, contrast, and natural impression of scenes when rendering high dynamic range images on conventional low dynamic range display devices. The method is simple and easy to implement. Our new approach uses a modified mapping function that takes into account the block-level log-average, as well as the log-average of the whole image. This allows for increased visibility in dark regions. For those images with sharp edges that may cause a halo effect, the halo effect can be reduced by using bilateral filtering when the block-level log-average is computed. Furthermore, the local details of the image can be enhanced by using a median filter. Experimental results show that the technique gives satisfactory results for a variety of images and avoids common artefacts, such as halos, gradient reversals, and the loss of local contrast. In particular, our method produces excellent results when an image has both dark and bright regions due to new block-level processing during tone mapping.

Single image dehazing via reliability guided fusion

Dark Channel Prior (DCP) accuracy is improved by block- and pixel-level processing.Reliability guided fusion of block and pixel dark channels produce fine transmission.DCP failure probability is reduced by an edge-preserving increase of the patch-size.DCP failure in sky is handled by limiting contrast boost of the sky-like regions.A downscaling method for fast transmission computation has also been proposed. This work addresses the shortcomings of the dark channel prior (DCP) and proposes a new and efficient method for transmission estimation. First, the accuracy of block-level and pixel-level dark channels is improved and a reliability map is generated. Then, through reliability guided fusion of block-level and pixel-level dark channels, a high-quality refined transmission map is obtained. The proposed method reduces the DCP failure probability and haloes by increasing the patch-size in an edge-preserving manner. DCP failure in the sky (bright) regions is handled by limiting the contrast boost of sky-like surfaces. This produces a more natural recovery of the sky regions. A downscaling method for fast transmission computation has also been proposed. Quantitative and qualitative comparisons show that the proposed method outperforms existing methods in terms of quality and speed. The proposed reliability guided fusion scheme is about 60 times faster than other well-known DCP based approaches.

Road vanishing point detection using weber adaptive local filter and salient-block-wise weighted soft voting

In this study, a novel and efficient technique is proposed for road vanishing point detection in challenging scenes. Currently, most existing texture-based methods detect the vanishing point using pixel wise texture orientation estimation and voting map generation, which suffers from high computational complexity. Since only road trails (e.g. road edges, ruts, and tire tracks) would contribute informative votes to vanishing point detection, the Weber adaptive local filter is proposed to distinguish road trails from background noise, which is envisioned to reduce the workload and to eliminate uninformative votes introduced by the background noise. Furthermore, instead of using the conventional pixel-wise voting scheme, the salient-block-wise weighted soft voting is developed to eliminate most of the noise votes introduced by incorrectly estimated pixel-wise texture orientations, and to further reduce the computation time of voting stage as well. The experimental results on the benchmark dataset demonstrate that the proposed method shows superior performance. The authors’ method is about ten times faster in detection speed and outperforms by 3.6% in detection accuracy, when compared with a well-known state-of-the-art approach.