scispace - formally typeset
Search or ask a question
Author

Xiahai Zhuang

Bio: Xiahai Zhuang is an academic researcher from Fudan University. The author has contributed to research in topics: Segmentation & Computer science. The author has an hindex of 28, co-authored 129 publications receiving 2770 citations. Previous affiliations of Xiahai Zhuang include Chinese Academy of Sciences & Shanghai Jiao Tong University.


Papers
More filters
Journal ArticleDOI
TL;DR: A whole heart segmentation method that employs multi-modality atlases from MRI and CT and adopts a new label fusion algorithm which is based on the proposed multi-scale patch (MSP) strategy and a new global atlas ranking scheme is presented for cardiac MRI.

348 citations

Journal ArticleDOI
TL;DR: This paper proposes a fully automatic whole heart segmentation framework based on two new image registration algorithms: the locally affine registration method (LARM) and the free-form deformations with adaptive control point status (ACPS FFDs).
Abstract: Magnetic resonance (MR) imaging has become a routine modality for the determination of patient cardiac morphology. The extraction of this information can be important for the development of new clinical applications as well as the planning and guidance of cardiac interventional procedures. To avoid inter- and intra-observer variability of manual delineation, it is highly desirable to develop an automatic technique for whole heart segmentation of cardiac magnetic resonance images. However, automating this process is complicated by the limited quality of acquired images and large shape variation of the heart between subjects. In this paper, we propose a fully automatic whole heart segmentation framework based on two new image registration algorithms: the locally affine registration method (LARM) and the free-form deformations with adaptive control point status (ACPS FFDs). LARM provides the correspondence of anatomical substructures such as the four chambers and great vessels of the heart, while the registration using ACPS FFDs refines the local details using a constrained optimization scheme. We validated our proposed segmentation framework on 37 cardiac MR volumes on the end-diastolic phase, displaying a wide diversity of morphology and pathology, and achieved a mean accuracy of 2.14 ± 0.63 mm (rms surface distance) and a maximal error of 4.31 mm.

253 citations

Book ChapterDOI
22 Sep 2013
TL;DR: A novel algorithm for the estimation of high-resolution cardiac MR images from single short-axis cardiac MR image stacks using an innovative super-resolution model which does not require explicit motion estimation and can be used for the reproducible estimation of 3D cardiac functional indices.
Abstract: The accurate measurement of 3D cardiac function is an important task in the analysis of cardiac magnetic resonance (MR) images. However, short-axis image acquisitions with thick slices are commonly used in clinical practice due to constraints of acquisition time, signal-to-noise ratio and patient compliance. In this situation, the estimation of a high-resolution image can provide an approximation of the underlaying 3D measurements. In this paper, we develop a novel algorithm for the estimation of high-resolution cardiac MR images from single short-axis cardiac MR image stacks. First, we propose to use a novel approximate global search approach to find patch correspondence between the short-axis MR image and a set of atlases. Then, we propose an innovative super-resolution model which does not require explicit motion estimation. Finally, we build an expectation-maximization framework to optimize the model. We validate the proposed approach using images from 19 subjects with 200 atlases and show that the proposed algorithm significantly outperforms conventional interpolation such as linear or B-spline interpolation. In addition, we show that the super-resolved images can be used for the reproducible estimation of 3D cardiac functional indices.

242 citations

Journal ArticleDOI
TL;DR: This work presents the methodologies and evaluation results for the WHS algorithms selected from the submissions to the Multi-Modality Whole Heart Segmentation (MM-WHS) challenge, in conjunction with MICCAI 2017.

216 citations

Journal ArticleDOI
Xiahai Zhuang1
TL;DR: In this article, the authors proposed a method for simultaneous registration and segmentation of multi-source images, using the multivariate mixture model (MvMM) and maximum of loglikelihood (LL) framework.
Abstract: The author proposes a method for simultaneous registration and segmentation of multi-source images, using the multivariate mixture model (MvMM) and maximum of log-likelihood (LL) framework. Specifically, the method is applied to the problem of myocardial segmentation combining the complementary information from multi-sequence (MS) cardiac magnetic resonance (CMR) images. For the image misalignment and incongruent data, the MvMM is formulated with transformations and is further generalized for dealing with the hetero-coverage multi-modality images (HC-MMIs). The segmentation of MvMM is performed in a virtual common space, to which all the images and misaligned slices are simultaneously registered. Furthermore, this common space can be divided into a number of sub-regions, each of which contains congruent data, thus the HC-MMIs can be modeled using a set of conventional MvMMs. Results show that MvMM obtained significantly better performance compared to the conventional approaches and demonstrated good potential for scar quantification as well as myocardial segmentation. The generalized MvMM has also demonstrated better robustness in the incongruent data, where some images may not fully cover the region of interest, and the full coverage can only be reconstructed combining the images from multiple sources.

176 citations


Cited by
More filters
Proceedings ArticleDOI
27 Jun 2016
TL;DR: This paper presents the first convolutional neural network capable of real-time SR of 1080p videos on a single K2 GPU and introduces an efficient sub-pixel convolution layer which learns an array of upscaling filters to upscale the final LR feature maps into the HR output.
Abstract: Recently, several models based on deep neural networks have achieved great success in terms of both reconstruction accuracy and computational performance for single image super-resolution. In these methods, the low resolution (LR) input image is upscaled to the high resolution (HR) space using a single filter, commonly bicubic interpolation, before reconstruction. This means that the super-resolution (SR) operation is performed in HR space. We demonstrate that this is sub-optimal and adds computational complexity. In this paper, we present the first convolutional neural network (CNN) capable of real-time SR of 1080p videos on a single K2 GPU. To achieve this, we propose a novel CNN architecture where the feature maps are extracted in the LR space. In addition, we introduce an efficient sub-pixel convolution layer which learns an array of upscaling filters to upscale the final LR feature maps into the HR output. By doing so, we effectively replace the handcrafted bicubic filter in the SR pipeline with more complex upscaling filters specifically trained for each feature map, whilst also reducing the computational complexity of the overall SR operation. We evaluate the proposed approach using images and videos from publicly available datasets and show that it performs significantly better (+0.15dB on Images and +0.39dB on Videos) and is an order of magnitude faster than previous CNN-based methods.

4,770 citations

Proceedings ArticleDOI
18 Jun 2018
TL;DR: This paper proposes residual dense block (RDB) to extract abundant local features via dense connected convolutional layers and uses global feature fusion in RDB to jointly and adaptively learn global hierarchical features in a holistic way.
Abstract: A very deep convolutional neural network (CNN) has recently achieved great success for image super-resolution (SR) and offered hierarchical features as well. However, most deep CNN based SR models do not make full use of the hierarchical features from the original low-resolution (LR) images, thereby achieving relatively-low performance. In this paper, we propose a novel residual dense network (RDN) to address this problem in image SR. We fully exploit the hierarchical features from all the convolutional layers. Specifically, we propose residual dense block (RDB) to extract abundant local features via dense connected convolutional layers. RDB further allows direct connections from the state of preceding RDB to all the layers of current RDB, leading to a contiguous memory (CM) mechanism. Local feature fusion in RDB is then used to adaptively learn more effective features from preceding and current local features and stabilizes the training of wider network. After fully obtaining dense local features, we use global feature fusion to jointly and adaptively learn global hierarchical features in a holistic way. Experiments on benchmark datasets with different degradation models show that our RDN achieves favorable performance against state-of-the-art methods.

2,860 citations

Reference EntryDOI
15 Oct 2004

2,118 citations

Posted Content
Yulun Zhang1, Kunpeng Li1, Kai Li1, Lichen Wang1, Bineng Zhong1, Yun Fu1 
TL;DR: This work proposes a residual in residual (RIR) structure to form very deep network, which consists of several residual groups with long skip connections, and proposes a channel attention mechanism to adaptively rescale channel-wise features by considering interdependencies among channels.
Abstract: Convolutional neural network (CNN) depth is of crucial importance for image super-resolution (SR). However, we observe that deeper networks for image SR are more difficult to train. The low-resolution inputs and features contain abundant low-frequency information, which is treated equally across channels, hence hindering the representational ability of CNNs. To solve these problems, we propose the very deep residual channel attention networks (RCAN). Specifically, we propose a residual in residual (RIR) structure to form very deep network, which consists of several residual groups with long skip connections. Each residual group contains some residual blocks with short skip connections. Meanwhile, RIR allows abundant low-frequency information to be bypassed through multiple skip connections, making the main network focus on learning high-frequency information. Furthermore, we propose a channel attention mechanism to adaptively rescale channel-wise features by considering interdependencies among channels. Extensive experiments show that our RCAN achieves better accuracy and visual improvements against state-of-the-art methods.

2,025 citations