Large Shape Variations Research Articles

BowelNet: Joint Semantic-Geometric Ensemble Learning for Bowel Segmentation From Both Partially and Fully Labeled CT Images.

Accurate bowel segmentation is essential for diagnosis and treatment of bowel cancers. Unfortunately, segmenting the entire bowel in CT images is quite challenging due to unclear boundary, large shape, size, and appearance variations, as well as diverse filling status within the bowel. In this paper, we present a novel two-stage framework, named BowelNet, to handle the challenging task of bowel segmentation in CT images, with two stages of 1) jointly localizing all types of the bowel, and 2) finely segmenting each type of the bowel. Specifically, in the first stage, we learn a unified localization network from both partially- and fully-labeled CT images to robustly detect all types of the bowel. To better capture unclear bowel boundary and learn complex bowel shapes, in the second stage, we propose to jointly learn semantic information (i.e., bowel segmentation mask) and geometric representations (i.e., bowel boundary and bowel skeleton) for fine bowel segmentation in a multi-task learning scheme. Moreover, we further propose to learn a meta segmentation network via pseudo labels to improve segmentation accuracy. By evaluating on a large abdominal CT dataset, our proposed BowelNet method can achieve Dice scores of 0.764, 0.848, 0.835, 0.774, and 0.824 in segmenting the duodenum, jejunum-ileum, colon, sigmoid, and rectum, respectively. These results demonstrate the effectiveness of our proposed BowelNet framework in segmenting the entire bowel from CT images.

Multivariate Concavity Amplitude Index (MCAI) for characterizing Heschl's gyrus shape

Heschl's gyrus (HG), which includes primary auditory cortex, is highly variable in its shape (i.e. gyrification patterns), between hemispheres and across individuals. Differences in HG shape have been observed in the context of phonetic learning skill and expertise, and of professional musicianship, among others. Two of the most common configurations of HG include single HG, where a single transverse temporal gyrus is present, and common stem duplications (CSD), where a sulcus intermedius (SI) arises from the lateral aspect of HG. Here we describe a new toolbox, called ‘Multivariate Concavity Amplitude Index’ (MCAI), which automatically assesses the shape of HG. MCAI works on the output of TASH, our first toolbox which automatically segments HG, and computes continuous indices of concavity, which arise when sulci are present, along the outer perimeter of an inflated representation of HG, in a directional manner. Thus, MCAI provides a multivariate measure of shape, which is reproducible and sensitive to small variations in shape. We applied MCAI to structural magnetic resonance imaging (MRI) data of N=181 participants, including professional and amateur musicians and from non-musicians. Former studies have shown large variations in HG shape in the former groups. We validated MCAI by showing high correlations between the dominant (i.e. highest) lateral concavity values and continuous visual assessments of the degree of lateral gyrification of the first gyrus. As an application of MCAI, we also replicated previous visually obtained findings showing a higher likelihood of bilateral CSDs in musicians. MCAI opens a wide range of applications in evaluating HG shape in the context of individual differences, expertise, disorder and genetics.

Spider mites (Acariformes, Tetranychidae) of the Massane Nature Reserve (France)

We present results of the inventory of the spider mites of the Massane Forest Nature Reserve made during several field samplings from 2013 to 2017. Nineteen species have been recorded, which represents more than one third of the known spider mite fauna in France. Among them, three new species were disclosed and are described: Bryobia (Bryobia) alberensis sp. nov. collected on Phyteuma spicatum and Helichrysum stoechas, B. (Periplanobia) polymorpha sp. nov. on Genista pilosa and Tetranychus visigothus sp. nov., on Eryngium campestre, respectively. In addition, three spider mite species, Bryobia (Periplanobia) serifiotica, Eotetranychus aurantii and Eotetranychus fagi are recorded from France for the first time and three plant species are recorded as new host plants for tetranychid mites. The species delineation of B. (P.) polymorpha sp. nov., that presented large variation in dorsal setae length and shape, was investigated using a combination of morphological and molecular data. This approach allowed us to conclude that, despite such variations in a morphological key character, all the specimens observed belong to the same species.

EARDS: EfficientNet and attention-based residual depth-wise separable convolution for joint OD and OC segmentation.

Glaucoma is the leading cause of irreversible vision loss. Accurate Optic Disc (OD) and Optic Cup (OC) segmentation is beneficial for glaucoma diagnosis. In recent years, deep learning has achieved remarkable performance in OD and OC segmentation. However, OC segmentation is more challenging than OD segmentation due to its large shape variability and cryptic boundaries that leads to performance degradation when applying the deep learning models to segment OC. Moreover, the OD and OC are segmented independently, or pre-requirement is necessary to extract the OD centered region with pre-processing procedures. In this paper, we suggest a one-stage network named EfficientNet and Attention-based Residual Depth-wise Separable Convolution (EARDS) for joint OD and OC segmentation. In EARDS, EfficientNet-b0 is regarded as an encoder to capture more effective boundary representations. To suppress irrelevant regions and highlight features of fine OD and OC regions, Attention Gate (AG) is incorporated into the skip connection. Also, Residual Depth-wise Separable Convolution (RDSC) block is developed to improve the segmentation performance and computational efficiency. Further, a novel decoder network is proposed by combining AG, RDSC block and Batch Normalization (BN) layer, which is utilized to eliminate the vanishing gradient problem and accelerate the convergence speed. Finally, the focal loss and dice loss as a weighted combination is designed to guide the network for accurate OD and OC segmentation. Extensive experimental results on the Drishti-GS and REFUGE datasets indicate that the proposed EARDS outperforms the state-of-the-art approaches. The code is available at https://github.com/M4cheal/EARDS.

A new vessel motion based method for parametric estimation of the waves encountered by the ship in a seaway

Inference of sea wave spectrum based on vessel motions records is a relatively cheap monitoring strategy that has potential to improve the vessel’s day-to-day operational efficiency and the safety of its crew. For this reason, several different approaches have been considered in recent years for turning this technique into practical tools. The methods available in the literature for implementing this strategy may be broadly categorized into parametric and non-parametric approaches. Apart from other aspects, they share the characteristics of generally performing the wave inference in the earth fixed reference frame and not in the ship’s own referential. This procedure has the obvious advantage of providing the absolute wave spectrum as the output, but it comes at the expense of a greater complexity of the estimation algorithm when dealing with the motions of a vessel with forward speed. In a large part, this complexity arises from the uncertainties in the sea state estimation associated with the well-known triple-value problem when dealing with waves reaching the ship from the stern. The present article proposes a new parametric approach for vessel motion-based wave estimation. In this approach, a parametric description of the encounter wave spectrum is presented, thus enabling the algorithm to execute the wave inference in the ship’s reference frame. For dealing with the large shape variability that is inherent to the encounter wave spectra, the proposed parametrization is based on the Weibull’s distribution mathematical function. A first verification of performance of the method is also provided for the case-study of an oceanographic vessel. At this point, such verification is restricted to model-scale tests comprising conditions with and without forward speed. The results obtained are promising and indicate that the methodology is indeed capable of providing good estimates of the waves encountered by the ship under different operating conditions.

TD-Net: A Hybrid End-to-End Network for Automatic Liver Tumor Segmentation From CT Images.

Liver tumor segmentation plays an essential role in diagnosis and treatment of hepatocellular carcinoma or metastasis. However, accurate and automatic tumor segmentation remains a challenging task, owing to vague boundaries and large variations in shapes, sizes, and locations of liver tumors. In this paper, we propose a novel hybrid end-to-end network, called TD-Net, which incorporates Transformer and direction information into convolution network to segment liver tumor from CT images automatically. The proposed TD-Net is composed of a shared encoder, two decoding branches, four skip connections, and a direction guidance block. The shared encoder is utilized to extract multi-level feature information, and the two decoding branches are respectively designed to produce initial segmentation map and direction information. To preserve spatial information, four skip connections are used to concatenate each encoder layer and its corresponding decoder layer, and in the fourth skip connection a Transformer module is constructed to extract global context. Furthermore, a direction guidance block is well-designed to rectify feature maps to further improve segmentation accuracy. Extensive experiments conducted on public LiTS and 3DIRCADb datasets validate that the proposed TD-Net can effectively segment liver tumor from CT images in an end-to-end manner and its segmentation accuracy surpasses those of many existing methods.

MDF-Net: A Multi-Scale Dynamic Fusion Network for Breast Tumor Segmentation of Ultrasound Images.

Breast tumor segmentation of ultrasound images provides valuable information of tumors for early detection and diagnosis. Accurate segmentation is challenging due to low image contrast between areas of interest; speckle noises, and large inter-subject variations in tumor shape and size. This paper proposes a novel Multi-scale Dynamic Fusion Network (MDF-Net) for breast ultrasound tumor segmentation. It employs a two-stage end-to-end architecture with a trunk sub-network for multiscale feature selection and a structurally optimized refinement sub-network for mitigating impairments such as noise and inter-subject variation via better feature exploration and fusion. The trunk network is extended from UNet++ with a simplified skip pathway structure to connect the features between adjacent scales. Moreover, deep supervision at all scales, instead of at the finest scale in UNet++, is proposed to extract more discriminative features and mitigate errors from speckle noise via a hybrid loss function. Unlike previous works, the first stage is linked to a loss function of the second stage so that both the preliminary segmentations and refinement subnetworks can be refined together at training. The refinement sub-network utilizes a structurally optimized MDF mechanism to integrate preliminary segmentation information (capturing general tumor shape and size) at coarse scales and explores inter-subject variation information at finer scales. Experimental results from two public datasets show that the proposed method achieves better Dice and other scores over state-of-the-art methods. Qualitative analysis also indicates that our proposed network is more robust to tumor size/shapes, speckle noise and heavy posterior shadows along tumor boundaries. An optional post-processing step is also proposed to facilitate users in mitigating segmentation artifacts. The efficiency of the proposed network is also illustrated on the "Electron Microscopy neural structures segmentation dataset". It outperforms a state-of-the-art algorithm based on UNet-2022 with simpler settings. This indicates the advantages of our MDF-Nets in other challenging image segmentation tasks with small to medium data sizes.

Predicting the Second-Order Nonlinear Optical Responses of Organic Materials: The Role of Dynamics.

The last 30 years have witnessed an ever-growing application of computational chemistry for rationalizing the nonlinear optical (NLO) responses of organic chromophores. More specifically, quantum chemical calculations proved highly helpful in gaining fundamental insights into the factors governing the magnitude and character of molecular first hyperpolarizabilities (β), be they either intrinsic to the chromophore molecular structure and arising from symmetry, chemical substitution, or π-electron delocalization, or induced by external contributions such as the laser probe or solvation and polarization effects. Most theoretical reports assumed a rigid picture of the investigated systems, the NLO responses being computed solely at the most stable geometry of the chromophores. Yet, recent developments combining classical molecular dynamics (MD) simulations and DFT calculations have evidenced the significant role of structural fluctuations, which may induce broad distributions of NLO responses, and even generate them in some instances.This Account presents recent case studies in which theoretical simulations have highlighted these effects. The discussion specifically focuses on the simulation of the second-order NLO properties that can be measured experimentally either from Hyper-Rayleigh Scattering (HRS) or Electric-Field Induced Second Harmonic Generation (EFISHG). More general but technical topics concerning several aspects of the calculations of hyperpolarizabilities are instead discussed in the Supporting Information.Selected examples include organic chromophores, photochromic systems, and ionic complexes in the liquid phase, for which the effects of explicit solvation, concentration, and chromophore aggregation are emphasized, as well as large flexible systems such as peptide chains and pyrimidine-based helical polymers, in which the relative variations of the responses were shown to be several times larger than their average values. The impact of geometrical fluctuations is also illustrated for supramolecular architectures with the examples of nanoparticles formed by organic dipolar dyes in water solution, whose soft nature allows for large shape variations translating into huge fluctuations in time of their NLO response, and of self-assembled monolayers (SAMs) based on indolino-oxazolidine or azobenzene switches, in which the geometrical distortions of the photochromic molecules, as well as their orientational and positional disorder within the SAMs, highly impact their NLO response and contrast upon switching. Finally, the effects of the rigidity and fluidity of the surrounding are evidenced for NLO dyes inserted in phospholipid bilayers.

Pyramidal Semantic Correspondence Networks

This paper presents a deep architecture, called pyramidal semantic correspondence networks (PSCNet), that estimates locally-varying affine transformation fields across semantically similar images. To deal with large appearance and shape variations that commonly exist among different instances within the same object category, we leverage a pyramidal model where the affine transformation fields are progressively estimated in a coarse-to-fine manner so that the smoothness constraint is naturally imposed. Different from the previous methods which directly estimate global or local deformations, our method first starts to estimate the transformation from an entire image and then progressively increases the degree of freedom of the transformation by dividing coarse cell into finer ones. To this end, we propose two spatial pyramid models by dividing an image in a form of quad-tree rectangles or into multiple semantic elements of an object. Additionally, to overcome the limitation of insufficient training data, a novel weakly-supervised training scheme is introduced that generates progressively evolving supervisions through the spatial pyramid models by leveraging a correspondence consistency across image pairs. Extensive experimental results on various benchmarks including TSS, Proposal Flow-WILLOW, Proposal Flow-PASCAL, Caltech-101, and SPair-71k demonstrate that the proposed method outperforms the lastest methods for dense semantic correspondence.

Morpho-pathogenic and molecular diversity among Rhizoctonia bataticola isolates causing dry root rot of chickpea in Madhya Pradesh, India

Aim: The present study was conducted to identify the morphological, pathological and molecular variations in isolates of Rhizoctonia bataticola, causing dry root rot of chickpea collected from different parts of Madhya Pradesh, India. Methodology: The isolation, purification and identification of Rhizoctonia bataticola isolates was done from dry root rot infected chickpea plants, collected from 23 districts of Madhya Pradesh, India. Total 32 isolates of pathogen were obtained and further used for morphological, cultural, pathological and molecular characterization. The acquired data were subjected to DARwin5 software analysis for variability studies. The DNA isolates of the pathogen were also obtained and analysed to study molecular variability by RAPD analysis and the bands produced were subjected to PAST software for cluster analysis. Results: Thirty two isolates of R. bataticola with diverse origin showed a large variation in shape, size and initiation time of sclerotia. These isolates were pathogenic and grouped into three categories; highly virulent (59%), moderately virulent (25%) and less virulent (16%), under artificially inoculated conditions in sick soil method. The study of pathogenic variability revealed that ICC12441, ICC11332 and Pusa 212 were susceptible to all the isolates evaluated, whereas ICC12450, Pusa362, Pusa1103 showed moderate resistance against maximum isolates. Based on molecular characterization, the isolates were grouped into four clusters indicating that no two or more isolates were similar. Interpretation: The isolates of R. bataticola of chickpea from different agro-climatic zones of Madhya Pradesh, possess variability in their morphological and cultural characteristics, which not limited to geographical boundaries. The present study would be extremely useful for dry root rot management, as well as, in identifying donors for resistance breeding programmes. Key words: Chickpea, Diversity, Dry root rot, Pathogenicity, Rhizoctonia bataticola

Explainable multi-module semantic guided attention based network for medical image segmentation

Automated segmentation of medical images is crucial for disease diagnosis and treatment planning. Medical image segmentation has been improved based on the convolutional neural networks (CNNs) models. Unfortunately, they are still limited by scenarios in which the segmentation objective has large variations in size, boundary, position, and shape. Moreover, current CNNs have low explainability, restricting their use in clinical decisions. In this paper, we involve substantial use of various attentions in a CNN model and present an explainable multi-module semantic guided attention based network (MSGA-Net) for explainable and highly accurate medical image segmentation, which involves considering the most significant spatial regions, boundaries, scales, and channels. Specifically, we present a multi-scale attention module (MSA) to extract the most salient features at various scales from medical images. Then, we propose a semantic region-guided attention mechanism (SRGA) including location attention (LAM), channel-wise attention (CWA), and edge attention (EA) modules to extract the most important spatial, channel-wise, boundary-related features for interested regions. Moreover, we present a sequence of fine-tuning steps with the SRGA module to gradually weight the significance of interesting regions while simultaneously reducing the noise. In this work, we experimented with three different types of medical images such as dermoscopic images (HAM10000 dataset), multi-organ CT images (CHAOS 2019 dataset), and Brain tumor MRI images (BraTS 2020 dataset). Extensive experiments on all types of medical images revealed that our proposed MSGA-Net substantially increased the overall performance of all metrics over the existing models. Moreover, displaying the attention feature maps has more explainability than state-of-the-art models.

A fine‐to‐coarse‐to‐fine weakly supervised framework for volumetric SD‐OCT image segmentation

AbstractObtaining accurate segmentation of central serous chorioretinopathy in spectral‐domain optical coherence tomography (SD‐OCT) is critical for the determination of the disease severity. Although existing methods achieve considerable segmentation results, they heavily depend on large‐scale data with high‐quality annotations. Also, the lesions bear a large shape variation across different patients, which are often difficult to encode. To address the above problems, we propose a fine‐to‐coarse‐to‐fine weakly supervised framework. Specifically, global alternate max‐avg pooling (GTP) network can be employed to locate the lesion regions accurately by using only image‐level annotations. A network module based on the GTP network and a semantic transfer module are proposed to iteratively guide the network to continuously discover and expand the target lesion regions. Then, we employ 3D grey distribution histogram to generate pseudo‐volumetric labels. Finally, a novel 3D level set loss function is proposed to perform coarse‐to‐fine volumetric segmentation. Experiments on a challenging dataset demonstrate that the performance of our proposed method is closer to those of models trained with pixel‐level supervision.

VarRCWA: An Adaptive High-Order Rigorous Coupled Wave Analysis Method.

Semianalytical methods, such as rigorous coupled wave analysis, have been pivotal in the numerical analysis of photonic structures. In comparison to other numerical methods, they have a much lower computational cost, especially for structures with constant cross-sectional shapes (such as metasurface units). However, when the cross-sectional shape varies even mildly (such as a taper), existing semianalytical methods suffer from high computational costs. We show that the existing methods can be viewed as a zeroth-order approximation with respect to the structure's cross-sectional variation. We derive a high-order perturbative expansion with respect to the cross-sectional variation. Based on this expansion, we propose a new semianalytical method that is fast to compute even in the presence of large cross-sectional shape variation. Furthermore, we design an algorithm that automatically discretizes the structure in a way that achieves a user-specified accuracy level while at the same time reducing the computational cost.

Real-time characterization of regimes between continuous-wave operation and mode locking in an all-normal dispersion ytterbium-doped fiber ring laser

In the present work, the temporal dynamics of an all-normal-dispersion ytterbium-doped fiber ring laser operating in regimes intermediate between continuous-wave operation and mode locking are studied experimentally. Exploiting the segmented memory data acquisition possibilities of an ultrafast digital oscilloscope and through the use of a specifically developed algorithm, the real-time waveform evolution anchored in absolute time could be retrieved, without relying on a fixed reference carried by the signal itself, which does not exist in these regimes. By controlling the time spacing between the successively acquired frames, several dynamics taking place over different time scales could be evidenced and described. These measurements highlight in particular the evolution cycles of intense localized structures including spikes that emerge, grow, decay, interact and describe peculiar trajectories in phase-space diagrams, whereas their temporal positions evolve, driven in particular by gain dynamics. In spite of their dramatic variability, these structures are found to be enduring features of these regimes. Analysis of these data helps assessing the degree of partial mode locking associated with these dynamics. The time-domain mapping technique optimized using segmented memory data acquisition is shown to be useful to characterize precisely highly dynamical evolutions such as those revealed in this work, which are dominated by structures that present large and complex variations in amplitude, shape and position, and develop over different time scales.

An Improved U-Net Segmentation Model That Integrates a Dual Attention Mechanism and a Residual Network for Transformer Oil Leakage Detection

Accurately detecting oil leakage from a power transformer is important to maintain its normal operation. Deep learning (DL) methods have achieved satisfactory performance in automatic oil detection, but challenges remain due to the small amount of training data and oil targets with large variations in position, shape, and scale. To manage these issues, we propose a dual attention residual U-net (DAttRes-Unet) within a U-net architecture that extensively uses a residual network as well as spatial and channel-wise attention modules. To overcome the vanishing gradient problem due to deeper layers and a small amount of training data, a residual module from ResNet18 is used to construct the encoder path in the U-net framework. Meanwhile, to overcome the issue of training difficulty for the network, inspired by the advantage of transfer learning, initial network parameters in the encoder are obtained from the pre-trained ResNet18 on the ImageNet dataset. Further, in the decoder path, spatial attention and channel attention are integrated to highlight oil-stained regions while suppressing the background or irrelevant parts/channels. To facilitate the acquisition of the fluorescence images of the transformer, we designed a portable acquisition device integrating an ultraviolet light source and a digital camera. The proposed network is trained on the amount of fluorescence images after data augmentation is used and tested on actual fluorescence images. The experiment results show that the proposed DAttRes-Unet network can recognize oil-stained regions with a high accuracy of 98.49% for various shapes and scales of oil leakage.

Multi-Scale Pathological Fluid Segmentation in OCT With a Novel Curvature Loss in Convolutional Neural Network.

The segmentation of pathological fluid lesions in optical coherence tomography (OCT), including intraretinal fluid, subretinal fluid, and pigment epithelial detachment, is of great importance for the diagnosis and treatment of various eye diseases such as neovascular age-related macular degeneration and diabetic macular edema. Although significant progress has been achieved with the rapid development of fully convolutional neural networks (FCN) in recent years, some important issues remain unsolved. First, pathological fluid lesions in OCT show large variations in location, size, and shape, imposing challenges on the design of FCN architecture. Second, fluid lesions should be continuous regions without holes inside. But the current architectures lack the capability to preserve the shape prior information. In this study, we introduce an FCN architecture for the simultaneous segmentation of three types of pathological fluid lesions in OCT. First, attention gate and spatial pyramid pooling modules are employed to improve the ability of the network to extract multi-scale objects. Then, we introduce a novel curvature regularization term in the loss function to incorporate shape prior information. The proposed method was extensively evaluated on public and clinical datasets with significantly improved performance compared with the state-of-the-art methods.

Automated cervical tumor segmentation on MR images using multi-view feature attention network

Precise cervical cancer treatment highly relies on accurate segmentation of cervical tumors from magnetic resonance (MR) images. However, this task is challenged by the inhomogeneous intensity distributions in MR images and the large variations in tumor shapes and locations. The large slice thickness further results in limited inter-slice correlations and 3D contextual information can be utilized, which influences the segmentation performance greatly. To tackle the above challenges and make full use of the 3D contextual information, a multi-view feature attention-based segmentation network (MVFA-Net) is proposed in this study. Toward the weak correlations among adjacent MR slices, features from different views of the volumetric MR images are extracted and treated individually and then fused by a channel-wise attention model. A cervical MR data set collected from 160 cervical cancer patients was employed to evaluate the performance of the proposed MVFA-Net. In comparison experiments, the proposed MVFA-Net outperforms the other eight medical image segmentation networks 2.6%-11.1% and 0.39 mm-0.97 mm on Dice similarity coefficient (DSC) and Average surface distance (ASD), respectively. Extensive ablation studies demonstrate the effectiveness of the proposed multi-view attention block and MVFA-Net. Additionally, with the trained segmentation network, no more than 6 s will be taken to segment one unseen patient, which is highly efficient for clinical practice. The presented segmentation network might be useful for cervical cancer treatment routine to improve the segmentation accuracy, consistency, and efficiency. The code is publicly available at: https://github.com/xyndameinv/MVFA-Net.

Three-dimensional affinity learning based multi-branch ensemble network for breast tumor segmentation in MRI

Accurate and automatic breast tumor segmentation based on dynamic contrast-enhancement magnetic resonance imaging (DCE-MRI) plays an important role in breast cancer analysis. However, the background parenchymal enhancement and large variations in tumor size, shape or appearance make the task very challenging, and also the segmentation performance is still not satisfactory, especially for non-mass enhancement (NME) and small size tumors (≤2 cm). To address these challenges, we propose a novel 3D affinity learning based multi-branch ensemble network for accurate breast tumor segmentation. Specifically, two different types of subnetworks are built to form a multi-branch network. The two subnetworks are equipped with effective operation components, i.e., residual connection and channel-wise attention or making use of dense connectivity patterns, which can process the input images in parallel. Second, we propose an end-to-end trainable 3D affinity learning based refinement module by calculating the similarities between features of voxels, which is useful in discovering more pixels belonging to breast tumors. Third, two local affinity matrices are constructed by 3D affinity learning, which are used to refine the segmentation outputs of two subnetworks, respectively. Finally, a novel ensemble module is proposed to combine the information derived from the subnetworks, which can hierarchically merge the local and global affinity matrices derived from subnetworks. A large-scale breast DCE-MR images dataset with 420 subjects are built for evaluation, and comprehensive experiments have been conducted to demonstrate that our proposed method achieves superior performance over state-of-the-art medical image segmentation methods.

Predicting Breast Tumor Malignancy Using Deep ConvNeXt Radiomics and Quality-Based Score Pooling in Ultrasound Sequences.

Breast cancer needs to be detected early to reduce mortality rate. Ultrasound imaging (US) could significantly enhance diagnosing cases with dense breasts. Most of the existing computer-aided diagnosis (CAD) systems employ a single ultrasound image for the breast tumor to extract features to classify it as benign or malignant. However, the accuracy of such CAD system is limited due to the large tumor size and shape variation, irregular and ambiguous tumor boundaries, and low signal-to-noise ratio in ultrasound images due to their noisy nature and the significant similarity between normal and abnormal tissues. To handle these issues, we propose a deep-learning-based radiomics method based on breast US sequences in this paper. The proposed approach involves three main components: radiomic features extraction based on a deep learning network, so-called ConvNeXt, a malignancy score pooling mechanism, and visual interpretations. Specifically, we employ the ConvNeXt network, a deep convolutional neural network (CNN) trained using the vision transformer style. We also propose an efficient pooling mechanism to fuse the malignancy scores of each breast US sequence frame based on image-quality statistics. The ablation study and experimental results demonstrate that our method achieves competitive results compared to other CNN-based methods.

A-LugSeg: Automatic and explainability-guided multi-site lung detection in chest X-ray images

Large variations in anatomical shape and size, too much overlap between anatomical structures, and inconsistent anatomical shapes make accurate lung segmentation in chest x-rays (CXR) a challenging problem. In this paper, we propose an automatic method called A-LugSeg that consists of two subnetworks for lung segmentation in CXRs. The first is a segmentation subnetwork based on a deep learning model (i.e., Mask-RCNN), which completes a coarse segmentation for each input CXR image. The second is a refinement subnetwork designed to optimize the coarse segmentation result by combining an improved closed principal curve method and an enhanced machine learning, where the lung contour’s explainability-guided mathematical model is expressed by the machine learning’s parameters. The performance of the proposed method is evaluated on three public datasets, namely the ShenZhen hospital Chest X-ray dataset (SZCX), Japanese Society of Radiological Technology dataset (JSRT), and Montgomery County chest x-ray dataset (MC), which contain the 662 CXRs, 247 CXRs, and 138 CXRs, respectively. We used different datasets for training/validation (SZCX) and testing (SZCX/JSRT/MC). Furthermore, we used six evaluation metrics to evaluate the performance of our proposed method, including Dice Similarity Coefficient (DSC), Jaccard Similarity Coefficient (Ω), Accuracy (ACC), Precision, Sensitivity, and Specificity. The obtained results (DSC = 0.973, Ω = 0.958, ACC = 0.972, and p-value for DSC < 0.001) for JSRT, (DSC = 0.971, Ω = 0.955, ACC = 0.97, and p-value for DSC < 0.001) for MC, (DSC = 0.972, Ω = 0.956, and ACC = 0.97) for hybrid datasets (JSRT + MC), and (Precision, Sensitivity, and Specificity are higher than 0.98) show the superior performance of the proposed dual subnetwork segmentation algorithm compared to the existing state of the art approaches.