Sungsoo Ha

Sungsoo Ha

Auto-generated profile, based on publicly available information

Sungsoo Ha

Computational Science Initiative, Brookhaven National Laboratory, Upton, New York 11973, USA

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 August 2017 in IEEE Transactions on Medical Imaging
Verified
Journal Article
Open access
SHSungsoo Ha
Sungsoo Ha
KMKlaus Mueller
Klaus Mueller
Iterative algorithms have become increasingly popular in computed tomography (CT) image reconstruction, since they better deal with the adverse image artifacts arising from low radiation dose image acquisition. But iterative methods remain computationally expensive. The main cost emerges in the projection and back projection operations, where accurate CT system modeling can greatly improve the quality of the reconstructed image. We present a framework that improves upon one particular aspect - the accurate projection of the image basis functions. It differs from current methods in that it substitutes the high computational complexity associated with accurate voxel projection by a small number of memory operations. Coefficients are computed in advance and stored in look-up tables parameterized by the CT system's projection geometry. The look-up tables only require a few kilobytes of storage and can be efficiently accelerated on the GPU. We demonstrate our framework with both numerical and clinical experiments and compare its performance with the current state-of-the-art scheme - the separable footprint method.

30 January 2013 in IEEE Transactions on Nuclear Science
Verified
Journal Article
Closed access
SHS. Ha
S. Ha
SMS. Matej
S. Matej
MIM. Ispiryan
M. Ispiryan
KMKlaus Mueller
Klaus Mueller
We describe a GPU-accelerated framework that efficiently models spatially (shift) variant system response kernels and performs forward- and back-projection operations with these kernels for the DIRECT (Direct Image Reconstruction for TOF) iterative reconstruction approach. Inherent challenges arise from the poor memory cache performance at non-axis aligned TOF directions. Focusing on the GPU memory access patterns, we utilize different kinds of GPU memory according to these patterns in order to maximize the memory cache performance. We also exploit the GPU instruction-level parallelism to efficiently hide long latencies from the memory operations. Our experiments indicate that our GPU implementation of the projection operators has slightly faster or approximately comparable time performance than FFT-based approaches using state-of-the-art FFTW routines. However, most importantly, our GPU framework can also efficiently handle any generic system response kernels, such as spatially symmetric and shift-variant as well as spatially asymmetric and shift-variant, both of which an FFT-based approach cannot cope with.

9 October 2012 in IEEE Transactions on Visualization and Computer Graphics
Verified
Journal Article
Open access
LKLars Kuhne
Lars Kuhne
JGJoachim Giesen
Joachim Giesen
ZZZhiyuan Zhang
Zhiyuan Zhang
SHSungsoo Ha
Sungsoo Ha
KMKlaus Mueller
Klaus Mueller
Color mapping and semitransparent layering play an important role in many visualization scenarios, such as information visualization and volume rendering. The combination of color and transparency is still dominated by standard alpha-compositing using the Porter-Duff over operator which can result in false colors with deceiving impact on the visualization. Other more advanced methods have also been proposed, but the problem is still far from being solved. Here we present an alternative to these existing methods specifically devised to avoid false colors and preserve visual depth ordering. Our approach is data driven and follows the recently formulated knowledge-assisted visualization (KAV) paradigm. Preference data, that have been gathered in web-based user surveys, are used to train a support-vector machine model for automatically predicting an optimized hue-preserving blending. We have applied the resulting model to both volume rendering and a specific information visualization technique, illustrative parallel coordinate plots. Comparative renderings show a significant improvement over previous approaches in the sense that false colors are completely removed and important properties such as depth ordering and blending vividness are better preserved. Due to the generality of the defined data-driven blending operator, it can be easily integrated also into other visualization frameworks.

12 February 2019 in Acta Crystallographica Section A Foundations and Advances
Verified
Journal Article
Open access
PDF file
XHXiaojing Huang
Xiaojing Huang
HYHanfei Yan
Hanfei Yan
YHYan He
Yan He
MGMingyuan Ge
Mingyuan Ge
HÖHande Öztürk
Hande Öztürk
YCYong S. Chu
Yong S. Chu
Multi-slice X-ray ptychography offers an approach to achieve images with a nanometre-scale resolution from samples with thicknesses larger than the depth of field of the imaging system by modeling a thick sample as a set of thin slices and accounting for the wavefront propagation effects within the specimen. Here, we present an experimental demonstration that resolves two layers of nanostructures separated by 500 nm along the axial direction, with sub-10 nm and sub-20 nm resolutions on two layers, respectively. Fluorescence maps are simultaneously measured in the multi-modality imaging scheme to assist in decoupling the mixture of low-spatial-frequency features across different slices. The enhanced axial sectioning capability using correlative signals obtained from multi-modality measurements demonstrates the great potential of the multi-slice ptychography method for investigating specimens with extended dimensions in 3D with high resolution.

9 May 2018 in IEEE Transactions on Computational Imaging
Verified
Journal Article
Closed access
SHSungsoo Ha
Sungsoo Ha
KMKlaus Mueller
Klaus Mueller
Statistical iterative reconstruction (SIR) algorithms have shown great potential for improving image quality in reduced and low dose X-Ray Computed Tomography (CT). However, high computational cost and long reconstruction times have so far prevented the use of SIR in practical applications. Various optimization algorithms have been proposed to make SIR parallelizable for execution on multi-core computational platforms, while others have sought to improve its convergence rate. Parallelizing on a set of decoupled voxels within an iterative coordinate descent (ICD) optimization framework has shown good promise to achieve both of these premises. However, so far these types of frameworks come at the price of additional complexities or are limited to parallel beam geometry only. We improve on this prior research and present a framework which also achieves parallelism by processing sets of independent voxels, but does not introduce additional complexities and has no restrictions on beam geometry. Our method uses a novel multi-voxel update (MVU) scheme within a general ICD framework fully optimized for acceleration on commodity GPUs. We also investigate different GPU memory access patterns to increase cache hit-rates that result in improved time performance in the ICD framework. Experiments demonstrate speedups of two orders of magnitude for clinical datasets in cone-beam CT geometry, compared to the single-voxel update (SVU) scheme native to conventional ICD-based SIR. Finally, since our MVU scheme operates on fully independent voxels it maintains the fast convergence properties of ICD-based SIR. Consequently, the speedups achieved by parallel computing are not diminished by slower convergence of the iterative updates or by any additional overhead to decouple conflicting voxels.