Jinhua Yu

Noise reduction and edge detection via kernel anisotropic diffusion

A novel kernel anisotropic diffusion (KAD) method is proposed for robust noise reduction and edge detection. The KAD incorporates a kernelized gradient operator in the diffusion, leading to more effective edge detection and providing a better control to the diffusion process. Adaptive diffusion threshold estimation and automatic diffusion termination criterion are also introduced to enhance the robustness of the KAD. The KAD outperforms several previous anisotropic diffusion-based methods for low SNR images.

Ultrasound speckle reduction by a SUSAN-controlled anisotropic diffusion method

An ultrasound speckle reduction method is proposed in this paper. The filter, which enhances the power of anisotropic diffusion with the Smallest Univalue Segment Assimilating Nucleus (SUSAN) edge detector, is referred to as the SUSAN-controlled anisotropic diffusion (SUSAN_AD). The SUSAN edge detector finds image features by using local information from a pseudo-global perspective. Thanks to the noise insensitivity and structure preservation properties of SUSAN, a better control can be provided to the subsequent diffusion process. To enhance the adaptability of the SUSAN_AD, the parameters of the SUSAN edge detector are calculated based on the statistics of a fully formed speckle (FFS) region. Different FFS estimation schemes are proposed for envelope-detected speckle images and log-compressed ultrasonic images. Adaptive diffusion threshold estimation and automatic diffusion termination criterion are employed to enhance the robustness of the method. Both synthetic and real ultrasound images are used to evaluate the proposed method. The performance of the SUSAN_AD is compared with four other existing speckle reduction methods. It is shown that the proposed method is superior to other methods in both noise reduction and detail preservation.

Noninvasive IDH1 mutation estimation based on a quantitative radiomics approach for grade II glioma

ObjectiveThe status of isocitrate dehydrogenase 1 (IDH1) is highly correlated with the development, treatment and prognosis of glioma. We explored a noninvasive method to reveal IDH1 status by using a quantitative radiomics approach for grade II glioma.MethodsA primary cohort consisting of 110 patients pathologically diagnosed with grade II glioma was retrospectively studied. The radiomics method developed in this paper includes image segmentation, high-throughput feature extraction, radiomics sequencing, feature selection and classification. Using the leave-one-out cross-validation (LOOCV) method, the classification result was compared with the real IDH1 situation from Sanger sequencing. Another independent validation cohort containing 30 patients was utilised to further test the method.ResultsA total of 671 high-throughput features were extracted and quantized. 110 features were selected by improved genetic algorithm. In LOOCV, the noninvasive IDH1 status estimation based on the proposed approach presented an estimation accuracy of 0.80, sensitivity of 0.83 and specificity of 0.74. Area under the receiver operating characteristic curve reached 0.86. Further validation on the independent cohort of 30 patients produced similar results.ConclusionsRadiomics is a potentially useful approach for estimating IDH1 mutation status noninvasively using conventional T2-FLAIR MRI images. The estimation accuracy could potentially be improved by using multiple imaging modalities.Key Points• Noninvasive IDH1 status estimation can be obtained with a radiomics approach.• Automatic and quantitative processes were established for noninvasive biomarker estimation.• High-throughput MRI features are highly correlated to IDH1 states.• Area under the ROC curve of the proposed estimation method reached 0.86.

Robust phase-based texture descriptor for classification of breast ultrasound images

BackgroundClassification of breast ultrasound (BUS) images is an important step in the computer-aided diagnosis (CAD) system for breast cancer. In this paper, a novel phase-based texture descriptor is proposed for efficient and robust classifiers to discriminate benign and malignant tumors in BUS images.MethodThe proposed descriptor, namely the phased congruency-based binary pattern (PCBP) is an oriented local texture descriptor that combines the phase congruency (PC) approach with the local binary pattern (LBP). The support vector machine (SVM) is further applied for the tumor classification. To verify the efficiency of the proposed PCBP texture descriptor, we compare the PCBP with other three state-of-art texture descriptors, and experiments are carried out on a BUS image database including 138 cases. The receiver operating characteristic (ROC) analysis is firstly performed and seven criteria are utilized to evaluate the classification performance using different texture descriptors. Then, in order to verify the robustness of the PCBP against illumination variations, we train the SVM classifier on texture features obtained from the original BUS images, and use this classifier to deal with the texture features extracted from BUS images with different illumination conditions (i.e., contrast-improved, gamma-corrected and histogram-equalized). The area under ROC curve (AUC) index is used as the figure of merit to evaluate the classification performances.Results and conclusionsThe proposed PCBP texture descriptor achieves the highest values (i.e. 0.894) and the least variations in respect of the AUC index, regardless of the gray-scale variations. It’s revealed in the experimental results that classifications of BUS images with the proposed PCBP texture descriptor are efficient and robust, which may be potentially useful for breast ultrasound CADs.

Deep Learning based Radiomics (DLR) and its usage in noninvasive IDH1 prediction for low grade glioma

Deep learning-based radiomics (DLR) was developed to extract deep information from multiple modalities of magnetic resonance (MR) images. The performance of DLR for predicting the mutation status of isocitrate dehydrogenase 1 (IDH1) was validated in a dataset of 151 patients with low-grade glioma. A modified convolutional neural network (CNN) structure with 6 convolutional layers and a fully connected layer with 4096 neurons was used to segment tumors. Instead of calculating image features from segmented images, as typically performed for normal radiomics approaches, image features were obtained by normalizing the information of the last convolutional layers of the CNN. Fisher vector was used to encode the CNN features from image slices of different sizes. High-throughput features with dimensionality greater than 1.6*104 were obtained from the CNN. Paired t-tests and F-scores were used to select CNN features that were able to discriminate IDH1. With the same dataset, the area under the operating characteristic curve (AUC) of the normal radiomics method was 86% for IDH1 estimation, whereas for DLR the AUC was 92%. The AUC of IDH1 estimation was further improved to 95% using DLR based on multiple-modality MR images. DLR could be a powerful way to extract deep information from medical images.

Subarray coherence based postfilter for eigenspace based minimum variance beamformer in ultrasound plane-wave imaging

This paper introduces a new beamformer, which combines the eigenspace based minimum variance (ESBMV) beamformer with a subarray coherence based postfilter (SCBP), for improving the quality of ultrasound plane-wave imaging. The ESBMV beamformer has been validated in improving the imaging contrast, but the difficulty in dividing the signal subspace limits the usage of it in the low signal-to-noise ratio (SNR) scenarios. Coherence factor (CF) based methods could optimize the output of a distortionless beamformer to reduce sidelobes, but the influence by the subarray decorrelation technique on the postfilter design has not attracted enough concern before. Accordingly, an ESBMV-SCBP beamformer was proposed in this paper, which used the coherence of the subarray signal to compute an SCBP to optimize the ESBMV results. Simulated and experimental data were used to evaluate the performance of the proposed method. The results showed that the ESBMV-SCBP method achieved an improved imaging quality compared with the ESBMV beamformer. In the simulation study, the contrast ratio (CR) for an anechoic cyst was improved by 9.88 dB and the contrast-to-noise ratio (CNR) was improved by 0.97 over the ESBMV. In the experimental study, the CR improvements for two anechoic cysts were 7.32 dB and 9.45 dB, while the CNRs were improved by 1.27 and 0.66, respectively. The ESBMV-SCBP also showed advantages over the ESBMV-Wiener beamformer in preserving a less grainy speckle, which is closer to that of distortionless beamformers and benefits the imaging contrast. With a relatively small extra computational load, the proposed method has potential to enhance the quality of the ultrasound plane-wave imaging.

Correspondence - Beam-domain eigenspace-based minimum variance beamformer for medical ultrasound imaging

The eigenspace-based minimum variance (ESBMV) beamformer can provide good imaging resolution and contrast; however, the performance is achieved at the cost of high computational complexity. In adaptive array processing, the beamspace method is an efficient way to lower the computational complexity. In this paper, we combine the beamspace method with the ESBMV beamformer and propose a beam-domain ESBMV beamformer. We demonstrate the feasibility of introducing the beamspace into the ESBMV beamformer and propose an effective method of forming the transform matrix based on the spatial spectrum of the array signals. We also illustrate the performance of the proposed beamformer when resolving point scatterers and a cyst phantom with both simulated and experimental data. The results show that the proposed method can achieve performance comparable to the ESBMV beamformer within much shorter time.

Object density-based image segmentation and its applications in biomedical image analysis

In many applications of medical image analysis, the density of an object is the most important feature for isolating an area of interest (image segmentation). In this research, an object density-based image segmentation methodology is developed, which incorporates intensity-based, edge-based and texture-based segmentation techniques. The proposed method consists of three main stages: preprocessing, object segmentation and final segmentation. Image enhancement, noise reduction and layer-of-interest extraction are several subtasks of preprocessing. Object segmentation utilizes a marker-controlled watershed technique to identify each object of interest (OI) from the background. A marker estimation method is proposed to minimize over-segmentation resulting from the watershed algorithm. Object segmentation provides an accurate density estimation of OI which is used to guide the subsequent segmentation steps. The final stage converts the distribution of OI into textural energy by using fractal dimension analysis. An energy-driven active contour procedure is designed to delineate the area with desired object density. Experimental results show that the proposed method is 98% accurate in segmenting synthetic images. Segmentation of microscopic images and ultrasound images shows the potential utility of the proposed method in different applications of medical image processing.

Plane wave compounding based on a joint transmitting-receiving adaptive beamformer

Plane wave compounding is a useful mode for ultrasound imaging because it can make a good compromise between imaging quality and frame rate. It is also useful for broad view ultrasound imaging. Traditional coherent plane wave compounding coherently sums the echo data of different steered transmitting waves as the output. The data correlation information of different emissions is not considered. Therefore, some adaptive techniques can be introduced into the compounding procedure. In this paper, we propose a Joint Transmitting-Receiving (JTR) adaptive beamforming scheme for plane wave compounding. Unlike traditional adaptive beamformers, the proposed beamforming scheme is designed for the 2-D data set obtained from multiple plane wave firings. It calculates both the transmitting aperture weights and the receiving aperture weights and then combines them into a 2-D adaptive weight function for compounding. Experiments are conducted on both simulated and phantom data. Results show that the proposed scheme has better performance on both point targets and cysts than the existing plane wave compounding approach. Because of the adaptive process in both apertures for compounding, an improved resolution is observed in both simulation and phantom studies. When the eigenanalysis is introduced, a contrast enhancement is achieved. For the simulated cyst, a contrast ratio (CR) improvement of 48% is achieved compared with the traditional plane wave compounding. For the phantom cyst, this improvement is 213.8%. The proposed scheme also has good robustness against sound velocity errors. Therefore, it is effective in enhancing the coherent plane wave compounding quality.

Anatomical location differences between mutated and wild-type isocitrate dehydrogenase 1 in low-grade gliomas.

ABSTRACT Anatomical location of gliomas has been considered as a factor implicating the contributions of a specific precursor cells during the tumor growth. Isocitrate dehydrogenase 1 (IDH1) is a pathognomonic biomarker with a significant impact on the development of gliomas and remarkable prognostic effect. The correlation between anatomical location of tumor and IDH1 states for low-grade gliomas was analyzed quantitatively in this study. Ninety-two patients diagnosed of low-grade glioma pathologically were recruited in this study, including 65 patients with IDH1-mutated glioma and 27 patients with wide-type IDH1. A convolutional neural network was designed to segment the tumor from three-dimensional magnetic resonance imaging images. Voxel-based lesion symptom mapping was then employed to study the tumor location distribution differences between gliomas with mutated and wild-type IDH1. In order to characterize the location differences quantitatively, the Automated Anatomical Labeling Atlas was used to partition the standard brain atlas into 116 anatomical volumes of interests (AVOIs). The percentages of tumors with different IDH1 states in 116 AVOIs were calculated and compared. Support vector machine and AdaBoost algorithms were used to estimate the IDH1 status based on the 116 location features of each patient. Experimental results proved that the quantitative tumor location measurement could be a very important group of imaging features in biomarker estimation based on radiomics analysis of glioma.