Automatic segmentation of gas plumes from multibeam water column images using a U-shape network

Cold seeps are widely developed on the seabed of continental margins and can form gas plumes due to the upward migration of methane-rich fluids.The detection and automatic segmentation of gas plumes are of great significance in locating and studying the cold seep system that is usually accompanied by hydrate layers in the subsurface.A multibeam echo-sounder system(MBES)can record the complete backscatter intensity of the water column,and it is one of the most effective means for detecting cold seeps.However,the gas plumes recorded in multibeam water column images(WCI)are usually blurred due to the interference of the complicated water environment and the sidelobes of the MBES,making it difficult to obtain the effective segmentation.Therefore,based on the existing UNet semantic segmentation network,this paper proposes an AP-UNet network combining the convolutional block attention module and the pyramid pooling module for the automatic segmentation and extraction of gas plumes.Comparative experiments are conducted among three traditional segmentation methods and two deep learning methods.The results show that the AP-UNet segmentation model can effectively suppress complicated water column noise interference.The segmentation precision,the Dice coefficient,and the recall rate of this model are 92.09%,92.00%,and 92.49%,respectively,which are 1.17%,2.10%,and 2.07%higher than the results of the UNet.

Automatic and Robust Segmentation of Multiple Sclerosis Lesions with Convolutional Neural Networks

The diagnosis of multiple sclerosis(MS)is based on accurate detection of lesions on magnetic resonance imaging(MRI)which also provides ongoing essential information about the progression and status of the disease.Manual detection of lesions is very time consuming and lacks accuracy.Most of the lesions are difficult to detect manually,especially within the grey matter.This paper proposes a novel and fully automated convolution neural network(CNN)approach to segment lesions.The proposed system consists of two 2D patchwise CNNs which can segment lesions more accurately and robustly.The first CNN network is implemented to segment lesions accurately,and the second network aims to reduce the false positives to increase efficiency.The system consists of two parallel convolutional pathways,where one pathway is concatenated to the second and at the end,the fully connected layer is replaced with CNN.Three routine MRI sequences T1-w,T2-w and FLAIR are used as input to the CNN,where FLAIR is used for segmentation because most lesions on MRI appear as bright regions and T1-w&T2-w are used to reduce MRI artifacts.We evaluated the proposed system on two challenge datasets that are publicly available from MICCAI and ISBI.Quantitative and qualitative evaluation has been performed with various metrics like false positive rate(FPR),true positive rate(TPR)and dice similarities,and were compared to current state-of-the-art methods.The proposed method shows consistent higher precision and sensitivity than other methods.The proposed method can accurately and robustly segment MS lesions from images produced by different MRI scanners,with a precision up to 90%.

Fully Automatic Scar Segmentation for Late Gadolinium Enhancement MRI Images in Left Ventricle with Myocardial Infarction

Numerous methods have been published to segment the infarct tissue in theleft ventricle, most of them either need manual work, post-processing, or suffer from poorreproducibility. We proposed an automatic segmentation method for segmenting the infarct tissue irleft ventricle with myocardial infarction. Cardiac images of a total of 60 diseased hearts (55 humanhearts and 5 porcine hearts) were used in this study. The epicardial and endocardial boundariesof the ventricles in every 2D slice of the cardiac magnetic resonance with late gadoliniumenhancement images were manually segmented. The subsequent pipeline of infarct tissuesegmentation is fully automatic. The segmentation results with the automatic algorithm proposed inthis paper were compared to the consensus ground truth. The median of Dice overlap between ourautomatic method and the consensus ground truth is 0.79. We also compared the automatic methodwith the consensus ground truth using different image sources from diferent centers with diferentscan parameters and different scan machines. The results showed that the Dice overlap with thepublic dataset was 0.83, and the overall Dice overlap was 0.79. The results show that our method isrobust with respect to different MRI image sources, which were scanned by different centers withdifferent image collection parameters. The segmentation accuracy we obtained is comparable toor better than that of the conventional semi-automatic methods. Our segmentation method may beuseful for processing large amount of dataset in clinic.

Automatic quantification of superficial foveal avascular zone in optical coherence tomography angiography implemented with deep learning

An accurate segmentation and quantification of the superficial foveal avascular zone(sFAZ)is important to facilitate the diagnosis and treatment of many retinal diseases,such as diabetic retinopathy and retinal vein occlusion.We proposed a method based on deep learning for the automatic segmentation and quantification of the sFAZ in optical coherence tomography angiography(OCTA)images with robustness to brightness and contrast(B/C)variations.A dataset of 405 OCTA images from 45 participants was acquired with Zeiss Cirrus HD-OCT 5000 and the ground truth(GT)was manually segmented subsequently.A deep learning network with an encoder–decoder architecture was created to classify each pixel into an sFAZ or non-sFAZ class.Subsequently,we applied largestconnected-region extraction and hole-filling to fine-tune the automatic segmentation results.A maximum mean dice similarity coefficient(DSC)of 0.976±0.011 was obtained when the automatic segmentation results were compared against the GT.The correlation coefficient between the area calculated from the automatic segmentation results and that calculated from the GT was 0.997.In all nine parameter groups with various brightness/contrast,all the DSCs of the proposed method were higher than 0.96.The proposed method achieved better performance in the sFAZ segmentation and quantification compared to two previously reported methods.In conclusion,we proposed and successfully verified an automatic sFAZ segmentation and quantification method based on deep learning with robustness to B/C variations.For clinical applications,this is an important progress in creating an automated segmentation and quantification applicable to clinical analysis.

Automated retinal layer segmentation on optical coherence tomography image by combination of structure interpolation and lateral mean filtering

Segmentation of layers in retinal images obtained by optical coherence tomography(OCT)has become an important clinical tool to diagnose ophthalmic diseases.However,due to the sus-ceptibility to speckle noise and shadow of blood vessels etc.,the layer segmentation technology based on a single image still fail to reach a satisfactory level.We propose a combination method of structure interpolation and lateral mean filtering(SI-LMF)to improve the signal-to-noise ratio based on one retinal image.Before performing one-dimensional lateral mean filtering to remove noise,structure interpolation was operated to eliminate thickness fluctuations.Then,we used boundary growth method to identify boundaries.Compared with existing segmentations,the method proposed in this paper requires less data and avoids the influence of microsaccade.The automatic segmentation method was verified on the spectral domain OCT volume images obtained from four normal objects,which successfully identified the boundaries of 10 physio-logical layers,consistent with the results based on the manual determination.

A mathematical morphological approach for region of interest coding of microscopy image compression

A novel mathematical morphological approach for region of interest(ROI) automatic determination and JPEG2000-based coding of microscopy image compression is presented.The algorithm is very fast and requires lower computing power,which is particularly suitable for some irregular region-based cell microscopy images with poor qualities.Firstly,an active threshold-based method is discussed to create a rough mask of regions of interest(cells).And then some morphological operations are designed and applied to achieve the segmentation of cells.In addition,an extra morphological operation,dilation,is applied to create the final mask with some redundancies to avoid the"edge effect"after removing false cells.Finally,ROI and region of background(ROB) are obtained and encoded individually in different compression ratio flexibly based on the JPEG2000,which can adjust the quality between ROI and ROB without coding for ROI shape.The experimental results certify the effectiveness of the proposed algorithm,and compared with JPEG2000,the proposed algorithm has better performance in both subjective quality and objective quality at the same compression ratios.

Automatic quantification of morphology on magnetic resonance images of the proximal tibia

The morphological quantification of the proximal tibia of the knee joint is important in knee replacement.Accurate knowledge of these parameters provides the basis for design of the tibial prosthesis and its fixation.Ideally,a prosthesis that is suitable for the morphological characteristics of Chinese knees is needed.In this paper,a deep learning automatic network framework is designed to achieve automatic segmentation and automatic quantitative analysis of magnetic resonance images of the tibia.An enhanced feature fusion network structure is designed,including high and low-level feature fusion path modules to create accurate segmentation of the tibia.A new method of extracting feature points and lines from outline contours of the proximal tibia is designed to automatically calculate six clinical morphological linear parameters of the tibia in real-time.The final result is an automatic visualisation of the tibial contour and automated extraction of tibial morphometric parameters.Validation of the results from our system against a gold standard obtained by manual processing by expert clinicians showed the Dice coefficient to be 0.97,the accuracy to be 0.98,and the correlation coefficients for all six morphological parameters of the automatic quantification of the tibia are above 0.96.The gender-specific study found that the values of the proximal tibial linear parameters of internal and external tibial diameter,anterior and posterior diameter,lateral plateau length,lateral plateau width,medial plateau length,and medial plateau width in male patients are significantly greater than in female patients(all P values<0.01).The results enrich the use of deep learning in medicine,providing orthopaedic specialists with a valuable and intelligent quantitative tool that can assess the progression and changes in osteoarthritis of the knee joint.

Automatic Video Segmentation Based on Information Centroid and Optimized SaliencyCut

We propose an automatic video segmentation method based on an optimized SaliencyCut equipped with information centroid(IC)detection according to level balance principle in physical theory.Unlike the existing methods,the image information of another dimension is provided by the IC to enhance the video segmentation accuracy.Specifically,our IC is implemented based on the information-level balance principle in the image,and denoted as the information pivot by aggregating all the image information to a point.To effectively enhance the saliency value of the target object and suppress the background area,we also combine the color and the coordinate information of the image in calculating the local IC and the global IC in the image.Then saliency maps for all frames in the video are calculated based on the detected IC.By applying IC smoothing to enhance the optimized saliency detection,we can further correct the unsatisfied saliency maps,where sharp variations of colors or motions may exist in complex videos.Finally,we obtain the segmentation results based on IC-based saliency maps and optimized SaliencyCut.Our method is evaluated on the DAVIS dataset,consisting of different kinds of challenging videos.Comparisons with the state-of-the-art methods are also conducted to evaluate our method.Convincing visual results and statistical comparisons demonstrate its advantages and robustness for automatic video segmentation.

Analysis of new bone,cartilage,and fibrosis tissue in healing murine allografts using whole slide imaging and a new automated histomorphometric algorithm

Histomorphometric analysis of histologic sections of normal and diseased bone samples,such as healing allografts and fractures,is widely used in bone research.However,the utility of traditional semi-automated methods is limited because they are labor-intensive and can have high interobserver variability depending upon the parameters being assessed,and primary data cannot be re-analyzed automatically.Automated histomorphometry has long been recognized as a solution for these issues,and recently has become more feasible with the development of digital whole slide imaging and computerized image analysis systems that can interact with digital slides.Here,we describe the development and validation of an automated application（algorithm）using Visiopharm＇s image analysis system to quantify newly formed bone,cartilage,and fibrous tissue in healing murine femoral allografts in high-quality digital images of H＆E/alcian blue-stained decalcified histologic sections.To validate this algorithm,we compared the results obtained independently using OsteoMeasureTM and Visiopharm image analysis systems.The intraclass correlation coefficient between Visiopharm and OsteoMeasure was very close to one for all tissue elements tested,indicating nearly perfect reproducibility across methods.This new algorithm represents an accurate and labor-efficient method to quantify bone,cartilage,and fibrous tissue in healing mouse allografts.

Development and validation of an endoscopic images-based deep learning model for detection with nasopharyngeal malignancies

Background:Due to the occult anatomic location of the nasopharynx and frequent presence of adenoid hyperpla-sia,the positive rate for malignancy identification during biopsy is low,thus leading to delayed or missed diagnosis for nasopharyngeal malignancies upon initial attempt.Here,we aimed to develop an artificial intelligence tool to detect nasopharyngeal malignancies under endoscopic examination based on deep learning.Methods:An endoscopic images-based nasopharyngeal malignancy detection model(eNPM-DM)consisting of a fully convolutional network based on the inception architecture was developed and fine-tuned using separate training and validation sets for both classification and segmentation.Briefly,a total of 28,966 qualified images were collected.Among these images,27,536 biopsy-proven images from 7951 individuals obtained from January 1st,2008,to December 31st,2016,were split into the training,validation and test sets at a ratio of 7:1:2 using simple randomiza-tion.Additionally,1430 images obtained from January 1st,2017,to March 31st,2017,were used as a prospective test set to compare the performance of the established model against oncologist evaluation.The dice similarity coef-ficient(DSC)was used to evaluate the efficiency of eNPM-DM in automatic segmentation of malignant area from the background of nasopharyngeal endoscopic images,by comparing automatic segmentation with manual segmenta-tion performed by the experts.Results:All images were histopathologically confirmed,and included 5713(19.7%)normal control,19,107(66.0%)nasopharyngeal carcinoma(NPC),335(1.2%)NPC and 3811(13.2%)benign diseases.The eNPM-DM attained an overall accuracy of 88.7%(95%confidence interval(CI)87.8%-89.5%)in detecting malignancies in the test set.In the prospective comparison phase,eNPM-DM outperformed the experts:the overall accuracy was 88.0%(95%CI 86.1%-89.6%)vs.80.5%(95%CI 77.0%-84.0%).The eNPM-DM required less time(40 s vs.110.0±5.8 min)and exhibited encouraging performance in automatic segmentation of nasopharyngeal malignant area from the background,with an average DSC of 0.78±0.24 and 0.75±0.26 in the test and prospective test sets,respectively.Conclusions:The eNPM-DM outperformed oncologist evaluation in diagnostic classification of nasopharyngeal mass into benign versus malignant,and realized automatic segmentation of malignant area from the background of nasopharyngeal endoscopic images.