MSF-GAN: Multi-Scale Fuzzy Generative Adversarial Network for Breast Ultrasound Image Segmentation

Automatic breast ultrasound image (BUS) segmentation is still a challenging task due to poor image quality and inherent speckle noise. In this paper, we propose a novel multi-scale fuzzy generative adversarial network (MSF-GAN) for breast ultrasound image segmentation. The proposed MSF-GAN consists of two networks: a generative network to generate segmentation maps for input BUS images, and a discriminative network that employs a multi-scale fuzzy (MSF) entropy module for discrimination. The major contribution of this paper is applying fuzzy logic and fuzzy entropy in the discriminative network which can distinguish the uncertainty of segmentation maps and groundtruth maps and forces the generative network to achieve better segmentation performance. We evaluate the performance of MSF-GAN on three BUS datasets and compare it with six state-of-the-art deep neural network-based methods in terms of five metrics. MSF-GAN achieves the highest mean IoU of 78.75%, 73.30%, and 71.12% on three datasets, respectively.

A Novel Adaptive Fuzzy Deep Learning Approach for Histopathologic Cancer Detection

We proposed a novel model that integrates the fuzzy theory and group equivariant convolutional neural network for histopathologic cancer detection. The proposed fuzzy group equivariant convolutional neural network consists of the convolutional network, a novel fuzzy global pooling layer, and a fully connected network. In the fuzzy global pooling layer, the generated feature maps are transferred into the fuzzy domain by two different fuzzification methods. One of the fuzzy feature maps exploits the uncertainty information of histopathologic images, and the other keeps the original information. Furthermore, the fuzzy feature maps are processed by using Min-max operations. The experiments verified that the proposed method could always find the maximum fuzzy entropy and exploit and present the uncertainty of histopathologic images well. The experiments using the benchmark dataset demonstrate that the proposed model becomes more accurate and outperforms the existing models including the benchmark models. Compared to the benchmark model with 89.8% of accuracy, 96.3% of AUC, and 0.260 of negative log-likelihood loss, the proposed model obtained 91.7% of accuracy, 97.2% of AUC, and 0.214 of negative log-likelihood loss.

Tumor saliency estimation for breast ultrasound images via breast anatomy modeling

Tumor saliency estimation aims to localize tumors by modeling the visual stimuli in medical images. However, it is a challenging task for breast ultrasound (BUS) image due to the complicated anatomic structure of the breast and poor image quality; and existing saliency estimation approaches only model the generic visual stimuli, e.g., local and global contrast, location, and feature correlation, and achieve poor performance for tumor saliency estimation. In this paper, we propose a novel optimization model to estimate tumor saliency by utilizing breast anatomy. First, we model breast anatomy and decompose breast ultrasound image into layers using Neutro-Connectedness; then utilize the layers to generate the foreground and background maps; and finally propose a novel objective function to estimate the tumor saliency by integrating the foreground map, background map, adaptive center bias, and region-based correlation cues. The extensive experiments demonstrate that the proposed approach obtains more accurate foreground and background maps with breast anatomy; especially, for the images having large or small tumors. Meanwhile, the new objective function can handle the images without tumors. The newly proposed method achieves state-of-the-art performance comparing to eight tumor saliency estimation approaches using two BUS datasets.

[Retrospective analysis of diagnosis and treatment of breast cancer in pregnancy]

Objective: To investigate the principles of diagnosis and treatment of breast cancer during pregnancy. Methods: Clinical data of patients with breast cancer during pregnancy admitted to Obstetrics and Gynecology Hospital of Fudan University between January 2012 to July 2017 were analyzed retrospectively. A total of 17 patients were diagnosed with breast cancer in pregnancy, the median age was 32 years (range from 25 to 45 years old), pathological staging revealed 2 patient with stage 0, 1 with stage Ⅱa, 7 with stage Ⅱb, 1 with stage Ⅲa, 2 with stage Ⅲc, 4 with stage Ⅳ. Results: Thirteen patients received surgical treatment in pregnancy, the gestational age at surgery was (27.7±4.6) weeks; 2 patients with ductal carcinoma in situ received mastectomy, 11 patients with breast cancer underwent modified radical mastectomy. In patients undergoing surgery during pregnancy, no prophylactic contractions were used in 4 patients who had been treated earlier, there were 2 patients with frequent contractions within 24 hours after operation in these patients. Follow-up 9 patients were given oral nifedipine to prevent contractions, no obvious contractions occurred after the operation. Seven patients received chemotherapy during pregnancy; the chemotherapy of 4 cases of triple negative breast cancer was weekly paclitaxel sequential epirubicin and cyclophosphamide, the chemotherapy of the other three patients was docetaxel sequential epirubicin and cyclophosphamide. Fifteen patients underwent cesarean section to terminate pregnancy, 2 patients underwent spontaneous labor. The gestational age of birth was (36.9 ±1.3) weeks. Less than 35 weeks of termination of pregnancy occurred in one patient, the fetus was delivered to the neonatal intensive care unit due to neonatal respiratory distress syndrome, and suffered from congenital dysaudia. The prognosis of the other 16 survived infants was good. The median follow-up time was 10 months (range from 4 to 27) months, in 13 patients of stage 0 to Ⅲc, one patient were diagnosed with bone metastasis at 12 months after surgery, the remaining 12 patients had no disease progression, the progression free survival rate was 12/13, the overall survival rate was 13/13. Among the 4 patients with stage Ⅳ, one died in 7 months after delivery, one had new liver metastasis in 8 months after delivery. The remaining 2 patients were in stable condition. Conclusions: Breast cancer in pregnancy can be treated effectively, multidisciplinary cooperation and detailed assessment of maternal-fetal risks and benefits are necessary. Chemotherapy during pregnancy is safe for maternal-fetal, but it needed a large sample of clinical studies and long-term follow-up. The neonatal outcome was associated with gestational age, and therefore premature delivery was avoided as much as possible during treatment.

[Analysis and discussion of risk factors related to dental implant failure]

The implant prosthesis has been extensively used in clinic recently, and implant failure is appearing. Many factors may cause the failure, and they work together generally. This paper summarizes and analyzes the failure cases related to implant treatment and relevant risk factors of oral implants in Department of Implantation, School of Stomatology, The Fourth Military Medical University during the past six years, in order to improve the success rate of implant prosthesis and provide guidance for clinical application.

An effective non-rigid registration approach for ultrasound image based on "demons" algorithm

Medical image registration is an important component of computer-aided diagnosis system in diagnostics, therapy planning, and guidance of surgery. Because of its low signal/noise ratio (SNR), ultrasound (US) image registration is a difficult task. In this paper, a fully automatic non-rigid image registration algorithm based on demons algorithm is proposed for registration of ultrasound images. In the proposed method, an "inertia force" derived from the local motion trend of pixels in a Moore neighborhood system is produced and integrated into optical flow equation to estimate the demons force, which is helpful to handle the speckle noise and preserve the geometric continuity of US images. In the experiment, a series of US images and several similarity measure metrics are utilized for evaluating the performance. The experimental results demonstrate that the proposed method can register ultrasound images efficiently, robust to noise, quickly and automatically.

An effective approach of lesion segmentation within the breast ultrasound image based on the cellular automata principle

In this paper, a novel lesion segmentation within breast ultrasound (BUS) image based on the cellular automata principle is proposed. Its energy transition function is formulated based on global image information difference and local image information difference using different energy transfer strategies. First, an energy decrease strategy is used for modeling the spatial relation information of pixels. For modeling global image information difference, a seed information comparison function is developed using an energy preserve strategy. Then, a texture information comparison function is proposed for considering local image difference in different regions, which is helpful for handling blurry boundaries. Moreover, two neighborhood systems (von Neumann and Moore neighborhood systems) are integrated as the evolution environment, and a similarity-based criterion is used for suppressing noise and reducing computation complexity. The proposed method was applied to 205 clinical BUS images for studying its characteristic and functionality, and several overlapping area error metrics and statistical evaluation methods are utilized for evaluating its performance. The experimental results demonstrate that the proposed method can handle BUS images with blurry boundaries and low contrast well and can segment breast lesions accurately and effectively.

Breast ultrasound image classification based on multiple-instance learning

Breast ultrasound (BUS) image segmentation is a very difficult task due to poor image quality and speckle noise. In this paper, local features extracted from roughly segmented regions of interest (ROIs) are used to describe breast tumors. The roughly segmented ROI is viewed as a bag. And subregions of the ROI are considered as the instances of the bag. Multiple-instance learning (MIL) method is more suitable for classifying breast tumors using BUS images. However, due to the complexity of BUS images, traditional MIL method is not applicable. In this paper, a novel MIL method is proposed for solving such task. First, a self-organizing map is used to map the instance space to the concept space. Then, we use the distribution of the instances of each bag in the concept space to construct the bag feature vector. Finally, a support vector machine is employed for classifying the tumors. The experimental results show that the proposed method can achieve better performance: the accuracy is 0.9107 and the area under receiver operator characteristic curve is 0.96 (p < 0.005).

Computer aided diagnosis system for breast cancer based on color Doppler flow imaging

Color Doppler flow imaging takes a great value in diagnosing and classifying benign and malignant breast lesions. However, scanning of color Doppler sonography is operator-dependent and ineffective. In this paper, a novel breast classification system based on B-Mode ultrasound and color Doppler flow imaging is proposed. First, different feature extraction methods were used to obtain the texture and geometric features from B-Mode ultrasound images. In color Doppler feature extraction stage, several spectrum features are extracted by applying blood flow velocity analysis to Doppler signals. Moreover, a velocity coherent vector method is proposed based on color coherence vector, which is helpful for designing to the optimize detection of flow indices from different blood flow velocity fields automatically. Finally, a support vector machine classifier with selected feature vectors is used to classify breast tumors into benign and malignant. The experimental results demonstrate that the proposed computer-aided diagnosis system is useful for reducing the unnecessary biopsy and death rate.

Completely automated segmentation approach for breast ultrasound images using multiple-domain features

Lesion segmentation is a challenging task for computer aided diagnosis systems. In this article, we propose a novel and fully automated segmentation approach for breast ultrasound (BUS) images. The major contributions of this work are: an efficient region-of-interest (ROI) generation method is developed and new features to characterize lesion boundaries are proposed. After a ROI is located automatically, two newly proposed lesion features (phase in max-energy orientation and radial distance), combined with a traditional intensity-and-texture feature, are utilized to detect the lesion by a trained artificial neural network. The proposed features are tested on a database of 120 images and the experimental results prove their strong distinguishing ability. Compared with other breast ultrasound segmentation methods, the proposed method improves the TP rate from 84.9% to 92.8%, similarity rate from 79.0% to 83.1% and reduces the FP rate from 14.1% to 12.0%, using the same database. In addition, sensitivity analysis demonstrates the robustness of the proposed method.

Scene Classification by Fuzzy Local Moments

The identification of images irrespective of their location, size and orientation is one of the important tasks in pattern analysis. The use of global moment features has been one of the most popular techniques for this purpose. We present a simple and effective method for gray-level image representation and identification which utilizes fuzzy radial moments of image segments (local moments) as features as opposed to global features. A multilayer perceptron neural network is employed for classification. Fuzzy entropy measure is applied to optimize the parameters of the membership function. The technique does not require translation, scaling or rotation of the image. Furthermore, it is suitable for parallel implementation which is an advantage for real-time applications. The classification capability and robustness of the technique are demonstrated by experiments on scaled, rotated and noisy gray-level images of uppercase and lowercase characters and digits of English alphabet, as well as the images of a set of tools. The proposed approach can handle rotation, scale and translation invariance, noise and fuzziness simultaneously.

Novel computer-aided diagnosis algorithms on ultrasound image: effects on solid breast masses discrimination

The objective of this study is to retrospectively investigate whether using the newly developed algorithms would improve radiologists' accuracy for discriminating malignant masses from benign ones on ultrasonographic (US) images. Five radiologists blinded to the histological results and clinical history independently interpreted 226 cases according to the sonographic lexicon of the fourth edition of the Breast Imaging Reporting and Data System and assigned a final assessment category to indicate the probability of malignancy. For each case, each radiologist provided three diagnoses: first with the original images, subsequently with the assistant of the resulting images processed by the proposed CAD algorithms which are called as processed images, and another using the processed images only. Observers' malignancy rating data were analyzed with the receiver operating characteristic (ROC) curve. For reading only with the processed images, areas under the ROC curve (A(z)) of each reader (0.863, 0.867, 0.859, 0.868, 0.878) were better than that with the original images (0.772, 0.807, 0.796, 0.828, 0.846), difference of the average A(z) between the twice reading was significant (p < 0.001). Compared with the results single used processed images, A(z) of utilizing the combined images were increased (0.866, 0.885, 0.872, 0.894, 0.903), but the difference is not statistically significant (p = 0.081). The proposed CAD method has potential to be a good aid to radiologists in distinguishing malignant breast solid masses from benign ones.

Automated segmentation of ultrasonic breast lesions using statistical texture classification and active contour based on probability distance

Because of its complicated structure, low signal/noise ratio, low contrast and blurry boundaries, fully automated segmentation of a breast ultrasound (BUS) image is a difficult task. In this paper, a novel segmentation method for BUS images without human intervention is proposed. Unlike most published approaches, the proposed method handles the segmentation problem by using a two-step strategy: ROI generation and ROI segmentation. First, a well-trained texture classifier categorizes the tissues into different classes, and the background knowledge rules are used for selecting the regions of interest (ROIs) from them. Second, a novel probability distance-based active contour model is applied for segmenting the ROIs and finding the accurate positions of the breast tumors. The active contour model combines both global statistical information and local edge information, using a level set approach. The proposed segmentation method was performed on 103 BUS images (48 benign and 55 malignant). To validate the performance, the results were compared with the corresponding tumor regions marked by an experienced radiologist. Three error metrics, true-positive ratio (TP), false-negative ratio (FN) and false-positive ratio (FP) were used for measuring the performance of the proposed method. The final results (TP = 91.31%, FN = 8.69% and FP = 7.26%) demonstrate that the proposed method can segment BUS images efficiently, quickly and automatically.

A novel approach to speckle reduction in ultrasound imaging

Speckle noise is inherent in ultrasound images, and it generally tends to reduce the resolution and contrast, thereby degrading the diagnostic accuracy of this modality. Speckle reduction is very important and critical for ultrasound imaging. In this paper, we propose a novel approach for speckle reduction using 2-D homogeneity and directional average filters. We have conducted experiments on numerous artificial images, clinic breast ultrasound images and vascular images. The experimental results are compared with that of other methods and the performance is evaluated using several merits, and they demonstrate that the proposed approach can reduce the speckle noise effectively without blurring the edges and damaging the textual information. It will be very useful for computer-aided diagnosis systems using ultrasound images.

Computerized-aid diagnosis of breast mass using ultrasound image

This paper presents a comparative study of the diagnostic results of the ultrasologists with/without using a novel enhancement algorithm for breast ultrasonic images based on fuzzy entropy principle and textural information. Totally, 350 ultrasound images of 115 cases were analyzed including 59 benign and 56 malignant lesions. The original breast images were fuzzified, the edge and textural information were extracted, and the images were enhanced. The original and enhanced images were assessed and evaluated by ultrasologists using double blind method before and after enhancement. The diagnostic sensitivity and specificity were calculated by the areas (Az) under the receiver operating characteristic (ROC) curves. And the two diagnostic results before and after enhancement were compared by Chi-square test in a 2 x 2 table. The results demonstrated that the discrimination rate of breast masses had been highly improved after employing the novel enhancement algorithm. The result indicates the sensitivity could be raised from 74.3% to 89.3% with the false-positive rate 14.3%, and the area (Az) under the ROC curve of diagnosis also increased from 0.84 to 0.93. The novel enhancement algorithm can increase the classification accuracy and decrease the rate of missing and misdiagnosis, and it is useful for breast cancer control.

Breast ultrasound image enhancement using fuzzy logic

Breast cancer is still a serious disease in the world. Early detection is very essential for breast cancer prevention and diagnosis. Breast ultrasound (US) imaging has been proven to be a valuable adjunct to mammography in the detection and classification of breast lesions. Because of the fuzzy and noisy nature of the US images and the low contrast between the breast cancer and tissue, it is difficult to provide an accurate and effective diagnosis. This paper presents a novel algorithm based on fuzzy logic that uses both the global and local information and has the ability to enhance the fine details of the US images while avoiding noise amplification and overenhancement. We normalize the images and then fuzzify the normalized images based on the maximum entropy principle. Edge and textural information are extracted to describe the lesion features and the scattering phenomenon of US images and the contrast ratio measuring the degree of enhancement is computed and modified. The defuzzification process is used to obtain the enhanced US images. To demonstrate the performance of the proposed approach, the algorithm was tested on 86 breast US images. Experimental results confirm that the proposed method can effectively enhance the details of the breast lesions without overenhancement or underenhancement.

Thresholding using two-dimensional histogram and fuzzy entropy principle

This paper presents a thresholding approach by performing fuzzy partition on a two-dimensional (2-D) histogram based on fuzzy relation and maximum fuzzy entropy principle. The experiments with various gray level and color images have demonstrated that the proposed approach outperforms the 2-D nonfuzzy approach and the one dimensional (1-D) fuzzy partition approach.

A hierarchical approach to color image segmentation using homogeneity

In this paper, a novel hierarchical approach to color image segmentation is studied. We extend the general idea of a histogram to the homogeneity domain. In the first phase of the segmentation, uniform regions are identified via multilevel thresholding on a homogeneity histogram. While we process the homogeneity histogram, both local and global information is taken into consideration. This is particularly helpful in taking care of small objects and local variation of color images. An efficient peak-finding algorithm is employed to identify the most significant peaks of the histogram. In the second phase, we perform histogram analysis on the color feature hue for each uniform region obtained in the first phase. We successfully remove about 99.7% singularity off the original images by redefining the hue values for the unstable points according to the local information. After the hierarchical segmentation is performed, a region merging process is employed to avoid over-segmentation. CIE(L*a*b*) color space is used to measure the color difference. Experimental results have demonstrated the effectiveness and superiority of the proposed method after an extensive set of color images was tested.

A novel approach to microcalcification detection using fuzzy logic technique

Breast cancer continues to be a significant public health problem in the United States. Approximately, 182,000 new cases of breast cancer are diagnosed and 46,000 women die of breast cancer each year. Even more disturbing is the fact that one out of eight women in the United States will develop breast cancer at some point during her lifetime. Since the cause of breast cancer remains unknown, primary prevention becomes impossible. Computer-aided mammography is an important and challenging task in automated diagnosis. It has great potential over traditional interpretation of film-screen mammography in terms of efficiency and accuracy. Microcalcifications are the earliest sign of breast carcinomas and their detection is one of the key issues for breast cancer control. In this study, a novel approach to microcalcification detection based on fuzzy logic technique is presented. Microcalcifications are first enhanced based on their brightness and nonuniformity. Then, the irrelevant breast structures are excluded by a curve detector. Finally, microcalcifications are located using an iterative threshold selection method. The shapes of microcalcifications are reconstructed and the isolated pixels are removed by employing the mathematical morphology technique. The essential idea of the proposed approach is to apply a fuzzified image of a mammogram to locate the suspicious regions and to interact the fuzzified image with the original image to preserve fidelity. The major advantage of the proposed method is its ability to detect microcalcifications even in very dense breast mammograms. A series of clinical mammograms are employed to test the proposed algorithm and the performance is evaluated by the free-response receiver operating characteristic curve. The experiments aptly show that the microcalcifications can be accurately detected even in very dense mammograms using the proposed approach.

Fuzzy homogeneity approach to multilevel thresholding

The spatial ambiguity among pixels has inherent vagueness rather than randomness, therefore, the conventional methods might not work well. We propose fuzzy homogeneity vectors to handle the greyness and spatial uncertainties among pixels, and to perform multilevel thresholding. The experimental results prove that the proposed approach works better than the histogram-based algorithms.

A nonlinear fluid model for pulmonary blood circulation

In this paper, we apply a quasi-one-dimensional unsteady nonlinear fluid model to study human pulmonary circulation. Eighteen generations of blood vessels composed of the branching arterial, capillary and venous distensible vessel segments make up the complete pulmonary circulation. The numerical result gives satisfactory agreement with the physiological experimental data: a dramatic pressure drop occurs in the arterioles and postcapillaries, a negative transmural pressure is shown in the postcapillary and small venous segments, a large reverse flow occurs in the main pulmonary artery during the diastolic period, and the reverse flow decreases gradually along the pulmonary tree. In the microgravity case where g = 180 cm s-2, the computation illustrated the effect of gravity force on the blood distribution in the different parts of the pulmonary circulation. The effect of gravity on the total output is not obvious. The effect of local factors which initiate the variations of the geometrical or pulmonary circulation can be stimulated quantitatively by this model. The proposal model can be very useful for clinical practice and for studying the extreme cases which are very difficult to investigate by experiments.

Space-Time Domain Expansion Approach to VLSI and Its Application to Hierarchical Scene Matching

VLSI technology has recently received increasing attention due to its high performance and high reliability. Designing a VLSI structure systematically for a given task becomes a very important problem to many computer engineers. In this paper, we present a method to transform a recursive computation task into a VLSI structure systematically. The main advantages of this approach are its simplicity and completeness. Several examples, such as vector inner product, matrix multiplication, convolution, comparison operations in relational database and fast Fourier transformation (FFT), are given to demonstrate the transformation procedure. Finally, we apply the proposed method to hierarchical scene matching. Scene matching refers to the process of locating or matching a region of an image with a corresponding region of another view of the same image taken from a different viewing angle or at a different time. We first present a constant threshold estimation for hierarchical scene matching. The VLSI implementation of the hierarchical scene matching is then described in detail.