Chunsoo Ahn

Fully automatic segmentation based on localizing active contour method

Cartilage segmentation is one of challenging issues because knee magnetic resonance (MR) images are consisted of thin sheet structure, intensity inhomogeneity, and low contrast between cartilage and muscle. In this paper, a fully automatic segmentation method for knee cartilage is proposed using spatial fuzzy c-mean clustering (SFCM) and morphological operators. The proposed method modifies the way to generate an approximate boundary of cartilage region, and combines it with localizing region-based active contour method, and overcomes limitations of previous methods. The performance of the proposed method is improved more than 10.8% by Dice similarity coefficient (DSC) in comparison with previous methods.

Dynamic frequency allocation of femtocells for guaranteed macrocell throughput in OFDMA networks

Recently, femtocells have been introduced for indoor coverage extension and higher bandwidth services. However, the frequency allocation problem between the femtocells and the macrocell causes co-channel interference to be treated carefully. In this paper, we propose a dynamic frequency resource allocation scheme using Fractional Frequency Reuse (FFR) while satisfying the required macrocell throughput with a higher priority. The proposed scheme guarantees target macrocell throughput and allocates resource to femtocells in a dynamic manner by using the macrocell base station information. Assumpting that the macrocell users have priority over femtocell users, the macrocell users are pre-allocated by the 3-sectored FFR. The macrocell base station then informs femtocells of the reused sub-carriers with the allocation information of the macrocell users. In this way, the interference between the macrocell users and femtocells is greatly reduced and the target macrocell throughput is fulfilled. The simulation results show that proposed scheme enhances both macrocell/femtocell throughputs, and satisfies macrocell throughput in the overall network.

Fast localised active contour for inhomogeneous image segmentation

The localised active contour framework has been widely used for image segmentation because it provides reliable results for inhomogeneous images. However, its computational complexity remains an issue. In this study, the authors introduce a fast algorithm based on the localised active contour framework. A key concept of the proposed algorithm is its consideration of the curve evolution based on the speed function only at active points that change across time, rather than at all points in a narrow band. This approach reduces computational time in the localised active contour. The authors additionally propose a modified speed function to address inhomogeneous image segmentation. The experimental results demonstrate significant advantages of the proposed method over existing methods, both in terms of computational efficiency and segmentation accuracy, for homogeneous and inhomogeneous images.

Dynamic Downlink Resource Management of Femtocells Using Power Control in OFDMA Networks

최근에 실내 커버리지 확장 및 광대역 데이터 서비스를 위한 가정용 기지국인 펨토셀에 대한 연구가 활발하다 . 펨토셀이 배치될 때, 기존에 있던 네트워크 상황 속에서 다양한 기술적인 문제들을 발생 시킬 수 있다 . 본 논문에서는 OFDMA 시스템에서 펨토 - 매크로셀 간의 간섭 완화에 대하여 연구하였고 , 동적인 하향링크 자원 관리 방법을 제안하였다. 펨토셀의 전송 파워를 매크로셀로부터 오는 신호의 세기에 기반을 두어서 선택하는 방법과 펨토셀 에게 할당할 부반송파를 동적으로 관리하는 방식을 제안하였다 . 이러한 방법으로 펨토셀과 매크로셀 사용자와의 간섭을 줄일 수 있다. 제안하는 방법은 수용율을 비롯한 여러 평가 지표로 성능을 검증하였다 .Key Words : Femtocell, Interference Management, OFDMA, Power Control, 펨토셀, 간섭관리, 직교 주파수 다중 분할방식, 전력제어ABSTRACT

Unsupervised segmentation of noisy and inhomogeneous images using global region statistics with non-convex regularization

Improving the segmentation of magnetic resonance (MR) images remains challenging because of the presence of noise and inhomogeneous intensity. In this paper, we present an unsupervised, multiphase segmentation model based on a Bayesian framework for both MR image segmentation and bias field correction in the presence of noise. In our model, global region statistics are utilized as segmentation criteria in order to classify regions with similar mean intensities but different variances. Additionally, we propose an edge indicator function based on a guided filter (instead of a Gaussian filter) that can preserve the underlying edges of the image obscured by noise. The proposed edge indicator function is integrated with non-convex regularization to overcome the influence of noise, resulting in more accurate segmentation. Furthermore, the proposed model utilizes a Markov random field to model the spatial correlation between neighboring pixels, which increases the robustness of the model under high-noise conditions. Experimental results demonstrate significant advantages in terms of both segmentation accuracy and bias field correction for inhomogeneous images in the presence of noise.

Knee Cartilage Defect Assessment using Cartilage Thickness Atlas

Abstract: Osteoarthritis is the most common chronic joint disease in the world. With its progression, cartilage thick-ness tends to diminish, which causes severe pain to human being. One way to examine the stage of osteoarthritisis to measure the cartilage thickness. When it comes to inter-subject study, however, it is not easy task to comparecartilage thickness since every human being has different cartilage structure. In this paper, we propose a methodto assess cartilage defect using MRI inter-subject thickness comparison. First, we used manual segmentation methodto build accurate atlas images and each segmented image was labeled as articular surface and bone-cartilage interfacein order to measure the thickness. Secondly, each point in the bone-cartilage interface was assigned the measuredthickness so that the thickness does not change after registration. We used affine transformation and SyGN to getdeformation fields which were then applied to thickness images to have cartilage thickness atlas. In this way, it ispossible to investigate pixel-by-pixel thickness comparison. Lastly, the atlas images were made according to theirosteoarthritis grade which indicates the degree of its progression. The result atlas images were compared using theanalysis of variance in order to verify the validity of our method. The result shows that a significant difference isexisted among them with p < 0.001.Key words: osteoarthritis, Kellgren & Lawrence grade, cartilage thickness, magnetic resonance imaging, inter-sub-ject study, analysis of variance

Cartilage-defect assessment by measuring thickness of knee MRI: Data from the osteoarthritis initiative

In this paper, we present a knee cartilage assessment technique to differentiate those in different osteoarthritis grades. We construct atlas for each grade and each atlas represents an average size of cartilage thickness. We build the atlas using manual segmentation of patellar cartilage following groupwise registration. We assign thickness value at the voxels in bone-cartilage interface in ahead. The thickness atlas is used for t-test to see how our method is accurate. T-test result shows our method finds significant difference between osteoarthritis grade 0 and 4 with p less than 0.001.

A performance comparison between expanding window fountain and hierarchical network codes in scalable video coding

Recently, network coding and channel coding are widely applied for scalable video coding in order to achieve the reliable transmission of multimedia data. Network coding as hierarchical network coding (HNC) provides unequal erasure protection for scalable layers and addresses the process of decoding layer by layer from higher to less importance. Similar with that scheme, expanding window fountain (EWF) code is designed for flexible and efficient solution of real-time scalable video multicast. In this paper, we investigate the performance of EWF coding and HNC for scalable video coding over lossy packet networks. The experimental results show that EWF coding achieves better performance than HNC in both good and bad channels.

A hybrid matching algorithm based on contour and motion information for depth estimation

Depth estimation is an important part of free viewpoint television (FTV) because accuracy of depth information affects directly the synthesized video quality at virtual viewpoint. However, generating an accurate depth map is a computationally complex process, which makes real-time implementation challenging. In order to obtain accurate depth information with low complexity, a hybrid matching technique is proposed. It is composed of three types of matching mode: pixel matching, 3x3 and 5x5 block matching. By using contour and motion information in matching mode selection, the proposed technique is more complied with human visual system which is more sensitive moving regions than static one. Experimental results show that the proposed algorithm can not only enhance the synthesized visual quality, but also reduce the complexity in matching process.

Control parameter setting of IEEE 802.11e for proportional throughput differentiation

IEEE 802.11e has been heavily researched with regard to support the Quality of Service (QoS) in wireless LAN environments The Enhanced Distributed Coordination Function (EDCF) mechanism of IEEE 802.11e has several control parameters for QoS provision, such as minimum/maximum contention window (CWmin/CWmax), retry limit, Arbitration Inter Frame Space (AIFS), and so on Through varying of these parameters, differentiated services can be provided to different priority classes, in terms of throughput, packet loss rate, and delay The parameters of the IEEE 802.11e EDCF mechanism that have the greatest effect on QoS are investigated in this paper Especially, Proportional Throughput Differentiation Service (PTDS) is proposed through mathematical analysis for proportional service differentiation between adjacent priority classes and proper setting of control parameters Finally, the minimum contention window size is confirmed to have the greatest effect with regard to throughput The validation of PTDS is validated with well-known network simulator, NS-2.