ABSTRACT
Image registration is of great clinical importance in computer aided diagnosis and surgical planning of liver diseases. Deep learning-based registration methods endow liver computed tomography (CT) image registration with characteristics of real-time and high accuracy. However, existing methods in registering images with large displacement and deformation are faced with the challenge of the texture information variation of the registered image, resulting in subsequent erroneous image processing and clinical diagnosis. To this end, a novel unsupervised registration method based on the texture filtering is proposed in this paper to realize liver CT image registration. Firstly, the texture filtering algorithm based on L0 gradient minimization eliminates the texture information of liver surface in CT images, so that the registration process can only refer to the spatial structure information of two images for registration, thus solving the problem of texture variation. Then, we adopt the cascaded network to register images with large displacement and large deformation, and progressively align the fixed image with the moving one in the spatial structure. In addition, a new registration metric, the histogram correlation coefficient, is proposed to measure the degree of texture variation after registration. Experimental results show that our proposed method achieves high registration accuracy, effectively solves the problem of texture variation in the cascaded network, and improves the registration performance in terms of spatial structure correspondence and anti-folding capability. Therefore, our method helps to improve the performance of medical image registration, and make the registration safely and reliably applied in the computer-aided diagnosis and surgical planning of liver diseases.
Subject(s)
Humans , Algorithms , Image Processing, Computer-Assisted , Liver Diseases , Tomography, X-Ray ComputedABSTRACT
Objective To assess alternations in left ventricular( LV) torsion parameters in healthy subjects and patients with premature ventricular complexes ( PVCs) from the right ventricular outflow tract (RVOT) .Methods ThirtypatientswithPVCsfromRVOTand31healthysubjectswereincluded.Two-dimensional speckle tracking imaging( 2D-STI) was applied to evaluate LV torsion parameters ,including LV rotational degrees in basal and apical levels respectively ,LV total torsional degrees ,and time to peak rotational and torsional degrees . All values of patients with PVCs were recorded during both sinus beats ( PVC-S) and premature ventricular beats ( PVC-V) . Results No significant difference was observed in left ventricular ejection fraction(LVEF) between PVC-S group and control subjects( P > 0 .05) ,while LV rotational degrees in apical levels[( 8 .47 ± 3 .54)° vs (9 .50 ± 3 .21)° , P = 0 .042] and LV total torsional degrees [ (11 .25 ± 6 .31)° vs (14 .00 ± 4 .07)° , P =0 .046] were significantly reduced in PVC-S group . In addition to the reduction of LV rotational degrees in apical levels[( 3 .93 ± 7 .23)° vs (9 .50 ± 3 .21)° , P =0 .000] and LV total torsional degrees[ ( 4 .35 ± 9 .62)° vs (14 .00 ± 4 .07)° , P =0 .000] ,lower apical levels[ ( -0 .57 ± 4 .44)° vs ( -5 .26 ± 3 .84)° , P =0 .000] and advanced LV rotational degrees [ ( 40 .5 ± 18 .6)％vs (48 .0 ± 9 .1)％ , P =0 .05] in basal levels were observed in PVC-V group in comparison with the control subjects . Compared with the PVC-S group ,PVC-V group showed lower LV rotational degrees in basal levels [ ( -0 .57 ± 4 .44)° vs ( -4 .57 ± 4 .57)° , P = 0 .000] and advanced time to peak rotational degrees [ (40 .5 ± 18 .6)％ vs (48 .1 ± 12 .6)％ , P =0 .018] ,as well as advanced time to peak and lower LV total torsionaldegrees[(39.3±15.4)％ vs(46.7±13.8)％ ,P =0.007 ;(4.35±9.62)°vs(11.25±6.31)°,P=0 .001] .Conclusions As to RVOT-PVC patients ,LV myocardial torsional motion has changed in PVC-S mainly manifested as a decrease of rotation degrees in apical levels even if the LVEF is still in the normal range . During PVC-V the rotation and twist degree is further reduced ,and the time sequence altered , accompanied with significantly decreased LVEF .
ABSTRACT
This paper studies the vein extraction technique based on the susceptibility weighted imaging (SWI) and introduced an improved self-adaptive threshold method based on the vessel enhancing diffusion. The approach employs the combination indicator of the local gray character, the global gray character and the tubular information of the vein. It first applies the vessel enhancing diffusion filter to enhance the continuity of the vein, increases the detection rate of tiny vein and suppresses the nucleus areas. And then it uses the improved self-adaptive threshold method to extract the vein. The results demonstrate that this approach can solve the problem above and extract the vein from the SWI image accurately.
Subject(s)
Humans , Cerebral Veins , Pathology , Diffusion Magnetic Resonance Imaging , Methods , Image Enhancement , Methods , Imaging, Three-DimensionalABSTRACT
Objective To compare CT perfusion changes of patients 1 month after living liver transplantation with healthy volunteers. Methods Twenty perioperative patients without complications and 25 healthy volunteers underwent dynamic multi-slice CT scan. On each targeted slice, liver was divided into 3 regions of interest (ROI) according to hepatic venous drainage, i.e. anterior segment area drained by middle hepatic veins (MHV) which was most likely congested, risk area drained by both MHV and RHV which was posterior to anterior segment, reference area drained by RHV. Perfusion parameters and relative perfusion difference were calculated.Results The relative perfusion differences of anterior segment and risk area in study group were as follows: BF 44.56±43.19 vs 25.41±27.49 ［ml/(min〖DK〗·100 g)］; BV 3.91±2.45 vs 2.15±1.86(ml/100 g); PVP 41.50±39.15 vs 22.66±23.82 ［ml/(min·100 g)］. Significant differences of RPDs of BF, BV and PVP were found in anterior segment (P<0.0001) and risk area (P＜0.01) compared with control group.Conclusion In perioperative period, there are hemodynamic differences in hepatic segments drained by RHV and MHV. BF, BV, PVP of anterior segment decreases obviously.
ABSTRACT
Objective To explore the difference of resting-state default-mode network functional connectivity in patients with treatment-resistant depression (TRD) and in healthy subjects. Methods Ten patients with TRD and 12 healthy control subjects underwent 440 s fMRI scans while resting quietly. Functional connectivity analysis was used to isolate the default mode network in each subject. Group maps of the default-mode network were generated and compared between the two groups. A within-group analysis was performed in the depressed group to explore effects of depression refractoriness on network functional connectivity. Results Functional connectivity of both side of middle temporal gyrus, rectal gyrus, precuneus gray matter, left orbital gyrus, right inferior parietal lobule, and post cingulate gyrus in TRD group weakened compared with that of the control subjects. Conclusion There are resting default network connection weakening in multiple brain areas in TRD patients, which may lead to self-control and emotional behavior abnormal in patients.