Seoung-Jun Oh

Fast terahertz reflection tomography using block-based compressed sensing

In this paper, a new fast terahertz reflection tomography is proposed using block-based compressed sensing. Since measuring the time-domain signal on two-dimensional grid requires excessive time, reducing measurement time is highly demanding in terahertz tomography. The proposed technique directly reduces the number of sampling points in the spatial domain without modulation or transformation of the signal. Compressed sensing in spatial domain suggests a block-based reconstruction, which substantially reduces computational time without degrading the image quality. An overlap-average method is proposed to remove the block artifact in the block-based compressed sensing. Fast terahertz reflection tomography using the block-based compressed sensing is demonstrated with an integrated circuit and parched anchovy as examples.

Blocking artifact reduction in JPEG-coded images

We proposed the post-processing method to reduce the blocking artifact in low-bit-rate coded JPEG images, which is based on minimum block-boundary-discontinuity criterion. We analyzed the statistical property of block boundary pixel difference, classified all blocks into one of two types using OSLD (overlapped sub-laplacian distribution), and compensated the quantization error depending on the type of block, adaptively. Our postprocessing method can provide better visual quality as well as PSNR than any other method.

Variable block size motion estimation implementation on compute unified device architecture (CUDA)

This paper proposes a highly parallel variable block size full search motion estimation algorithm with concurrent parallel reduction (CPR) on graphics processing unit (GPU) using compute unified device architecture (CUDA). This approach minimizes memory access latency by using high-speed on-chip memory of GPU. By applying parallel reductions concurrently depending on the amount of data and the data dependency, the proposed approach increases thread utilization and decreases the number of synchronization points which cause latency. Experimental results show that the proposed approach achieves substantial improvement up to 92 times than the central processing unit (CPU) only counterpart.

Fast motion estimation for HEVC with adaptive search range decision on CPU and GPU

In this paper, we propose a fast Motion Estimation (ME) algorithm with Adaptive Search Range (ASR) decision to further accelerate the Graphics Processing Units (GPU)-based ME for High Efficiency Video Coding (HEVC). The proposed approach adaptively decides search ranges on the Central Processing Unit (CPU) and transfers them to the GPU. Then, the GPU performs ME process in parallel. The proposed approach solves the dependency problem in the Motion Vector Predictor (MVP) derivation stage by using only temporal Motion Vectors (MVs). The proposed algorithm yields the total encoding time reduction of 40.3% with negligible Rate Distortion (RD) loss of 1.2%. In terms of ME, the GPU-based ME with ASR decision provides the time reduction of 54.7% and 1446.8× speed-up on average compared to the GPU-based ME without ASR decision and the full-search ME in the reference model, respectively.

A novel Video Quality impairment Monitoring Scheme over an IPTY service with packet loss

In this paper, we propose a novel video quality impairment monitoring scheme, called VR-VQMS (Visual Rhythm based Video Quality Monitoring Scheme), over an IPTV service in the presence of reduced reference. The proposed scheme quantifies the amount of quality impairment due to packet loss experienced by end users. Visual rhythm is used as feature information in the reduced reference method. We define NPSNR (Networked Peak-to-peak Signal-to-Noise Ratio) as a new objective quality metric. We also define several additional objective and subjective metrics including NMOS (Networked Mean Opinion Score) to obtain statistics about location, duration, number of times, and amount of quality impairment. Two scenarios, on-line and off-line VR-VQMS, are proposed as the simple yet practical solutions for a commercial IPTV service. Simulation results show that the proposed scheme closely approximates the results from the full reference method and gives good estimation of subjective quality experienced by test observers. We expect that the proposed scheme can be implemented as a small and light agent program running on a resource-limited set-top box.

An efficient QoS negotiation mechanism for IPTV service in heterogeneous networks

IPTV service has to guarantee the quality of service, like the existing broadcasting service. Providing QoS guarantee is an essential condition for IPTV service. To support the QoS-guaranteed IPTV service, the service over best-effort networks should consider both delay and packet loss caused by network congestion. With the rapid growth of wireless access networks, IPTV service should also consider the heterogeneous networks and terminals among the end users. In this paper, we propose a novel QoS negotiation mechanism for IPTV service that includes the quality negotiation function and uses the SIP. The quality negotiation function appropriately adapts the quality level of IPTV stream to user environments such as terminal capability and network characteristic. Through the simulation, we prove that our QoS negotiation mechanism appropriately adjusts the quality of IPTV stream.

Fast motion estimation for HEVC on graphics processing unit (GPU)

AbstractnThe recent video compression standard, HEVC (high efficiency video coding), will most likely be used in various applications in the near future. However, the encoding process is far too slow for real-time applications. At the same time, computing capabilities of GPUs (graphics processing units) have become more powerful in these days. In this paper, we have proposed a GPU-based parallel motion estimation (ME) algorithm to enhance the performance of an HEVC encoder. A frame is partitioned into two subframes for pipelined execution to improve GPU utilization. The flow chart is redetermined to solve data hazards in the pipelined execution. Two new methods are introduced in the proposed ME: decision of a representative search center position (RSCP) and warp-based concurrent parallel reduction (WCPR). A RSCP employs motion vectors of a co-located CTU in a previously encoded frame to solve a dependency problem in parallel computation with negligible coding loss. WCPR concurrently executes several parallel reduction operations, which increases the thread utilization from 20 to 89xa0% without any thread synchronization. The proposed encoder can make the portion of ME in the encoder negligible with 2.2xa0% bitrate increase against the HEVC test model (HM) encoder. In terms of ME, the proposed ME is 130.7 times faster than that of the HM encoder.

A novel video copy detection method based on statistical analysis

The careless illegally copied contents have been rising serious social problem as internet and multimedia technologies are developing. Therefore, effective and efficient copy detection techniques are required for content management and rights protection. In this paper, we propose a content-based hierarchical video copy detection method that estimates similarity using statistical characteristics between an original video and its spatial variations. We rank luminance average value of video that is robust to the special transformation, and choose similar video sequences named as candidate segments in huge amount of database to reduce processing time and complexity. Finally, we detect the copied video by using the hypothesis test of mean. As experiment result, proposed method has similar copy detection ratio accuracy to the reference method while our processing time and complexity are less than those of the reference since we can reduce the number of keyframes up to 50%. Also, the proposed method can efficiently detect spatial variations such as blur, contrast change, zoom in, and zoom out.

Residual signal compression based on the blind signal decomposition for video coding

In this paper, a new residual signal compression method is proposed based on the blind signal decomposition for video coding. Blind signal decomposition is derived based on the fact that most of the natural signals in the real world could be decomposed into their basis signals and their weight values used in the composition process. In the proposed video coding system, composite data generated by adding two or more blocks are coded. Then the proposed decoder parses the coded bitstream and reconstructs the composite residual. The reconstructed composite residual is decomposed into the original residual blocks based on the blind signal decomposition. In the proposed system, the blind source separation is selectively used, depending on the performance of source separation. It is found that we can achieve approximately 2 ∼ 3dB gain by embedding our algorithm into an MPEG-4 baseline encoder.

A Priority-based Differentiated Streaming scheme to guarantee the media quality in mobile WiMAX

Mobile WiMAX can provide high-speed wireless access to mobile users. It also supports applications and services with diverse QoS (Quality of Service) requirements. However, it does not guarantee media quality for multimedia streaming services using scalable video coding schemes, because it does not consider essential characteristics of video coding techniques and network congestion. To solve this problem, we propose a new differentiated streaming scheme in mobile WiMAX, called P-DS (Priority-based Differentiated Streaming), which uses the differentiated transmission of media frames according to network congestion and wireless channel status. The proposed P-DS scheme guarantees high quality of multimedia services through video decoding reliability by transmitting differentiated media frames when network congestion and wireless channel errors occur.