Category: Nevin Manimala

IEEE Trans Med Imaging. 2022 Nov 29;PP. doi: 10.1109/TMI.2022.3225422. Online ahead of print.

ABSTRACT

Anomaly detection in fundus images remains challenging due to the fact that fundus images often contain diverse types of lesions with various properties in locations, sizes, shapes, and colors. Current methods achieve anomaly detection mainly through reconstructing or separating the fundus image background from a fundus image under the guidance of a set of normal fundus images. The reconstruction methods, however, ignore the constraint from lesions. The separation methods primarily model the diverse lesions with pixel-based independent and identical distributed (i.i.d.) properties, neglecting the individualized variations of different types of lesions and their structural properties. And hence, these methods may have difficulty to well distinguish lesions from fundus image backgrounds especially with the normal personalized variations (NPV). To address these challenges, we propose a patch-based non-i.i.d. mixture of Gaussian (MoG) to model diverse lesions for adapting to their statistical distribution variations in different fundus images and their patch-like structural properties. Further, we particularly introduce the weighted Schatten p-norm as the metric of low-rank decomposition for enhancing the accuracy of the learned fundus image backgrounds and reducing false-positives caused by NPV. With the individualized modeling of the diverse lesions and the background learning, fundus image backgrounds and NPV are finely learned and subsequently distinguished from diverse lesions, to ultimately improve the anomaly detection. The proposed method is evaluated on two real-world databases and one artificial database, outperforming the state-of-the-art methods.

PMID:36446017 | DOI:10.1109/TMI.2022.3225422

IEEE Trans Image Process. 2022 Nov 29;PP. doi: 10.1109/TIP.2022.3223793. Online ahead of print.

ABSTRACT

This study presents a high-accuracy, efficient, and physically induced method for 3D point cloud registration, which is the core of many important 3D vision problems. In contrast to existing physics-based methods that merely consider spatial point information and ignore surface geometry, we explore geometry aware rigid-body dynamics to regulate the particle (point) motion, which results in more precise and robust registration. Our proposed method consists of four major modules. First, we leverage the graph signal processing (GSP) framework to define a new signature, i.e., point response intensity for each point, by which we succeed in describing the local surface variation, resampling keypoints, and distinguishing different particles. Then, to address the shortcomings of current physics-based approaches that are sensitive to outliers, we accommodate the defined point response intensity to median absolute deviation (MAD) in robust statistics and adopt the X84 principle for adaptive outlier depression, ensuring a robust and stable registration. Subsequently, we propose a novel geometric invariant under rigid transformations to incorporate higher-order features of point clouds, which is further embedded for force modeling to guide the correspondence between pairwise scans credibly. Finally, we introduce an adaptive simulated annealing (ASA) method to search for the global optimum and substantially accelerate the registration process. We perform comprehensive experiments to evaluate the proposed method on various datasets captured from range scanners to LiDAR. Results demonstrate that our proposed method outperforms representative state-of-the-art approaches in terms of accuracy and is more suitable for registering large-scale point clouds. Furthermore, it is considerably faster and more robust than most competitors. [Our implementation will be released at https://github.com/zikai1/GraphReg].

PMID:36446012 | DOI:10.1109/TIP.2022.3223793

IEEE Trans Cybern. 2022 Nov 29;PP. doi: 10.1109/TCYB.2022.3221544. Online ahead of print.

ABSTRACT

Convolutional neural networks (CNNs) have attracted much research attention and achieved great improvements in single-image dehazing. However, previous learning-based dehazing methods are mainly trained on synthetic data, which greatly degrades their generalization capability on natural hazy images. To address this issue, this article proposes a semi-supervised learning approach for single-image dehazing, where both synthetic and realistic images are leveraged during training. Considering the situation that it is hard to obtain the realistic pairs of hazy and haze-free images, how to utilize the realistic data is not a trivial work. In this article, a domain alignment module is introduced to narrow the distribution distance between synthetic data and realistic hazy images in a latent feature space. Meanwhile, a haze-aware attention module is designed to describe haze densities of different regions in the image, thus adaptively responds for different hazy areas. Furthermore, the dark channel prior is introduced to the framework to improve the quality of the unsupervised learning results by considering the statistical characters of haze-free images. Such a semi-supervised design can significantly address the domain shift issue between the synthetic and realistic data, and improve generalization performance in the real world. Experiments indicate that the proposed method obtains state-of-the-art performance on both public synthetic and realistic hazy images with better visual results.

PMID:36445999 | DOI:10.1109/TCYB.2022.3221544

IEEE Trans Pattern Anal Mach Intell. 2022 Nov 29;PP. doi: 10.1109/TPAMI.2022.3225418. Online ahead of print.

ABSTRACT

This paper presents a generic probabilistic framework for estimating the statistical dependency and finding the anatomical correspondences among an arbitrary number of medical images. The method builds on a novel formulation of the N-dimensional joint intensity distribution by representing the common anatomy as latent variables and estimating the appearance model with nonparametric estimators. Through connection to maximum likelihood and the expectation-maximization algorithm, an information-theoretic metric called X-metric and a co-registration algorithm named X-CoReg are induced, allowing groupwise registration of the N observed images with computational complexity of O(N). Moreover, the method naturally extends for a weakly-supervised scenario where anatomical labels of certain images are provided. This leads to a combined-computing framework implemented with deep learning, which performs registration and segmentation simultaneously and collaboratively in an end-to-end fashion. Extensive experiments were conducted to demonstrate the versatility and applicability of our model, including multimodal groupwise registration, motion correction for dynamic contrast enhanced magnetic resonance images, and deep combined computing for multimodal medical images. Results show the superiority of our method in various applications in terms of both accuracy and efficiency, highlighting the advantage of the proposed representation of the imaging process.

PMID:36445992 | DOI:10.1109/TPAMI.2022.3225418

Am Surg. 2022 Nov 29:31348221142590. doi: 10.1177/00031348221142590. Online ahead of print.

ABSTRACT

BACKGROUND: The Enhanced Recovery After Surgery (ERAS) society lists early mobilization as one of their recommendations for improving patient outcomes following colorectal surgery. The level of supporting evidence, however, is relatively weak, and furthermore, the ERAS guidelines do not clearly define “early” mobilization. In this study, we define mobilization in terms of time to first ambulation after surgery and develop an outcome-based cutoff for early mobilization.

METHODS: This is a retrospective cohort study comprised of 291 patients who underwent colorectal operations at a large, academic medical center from June to December 2019. Three cutoffs (12 hours, 24 hours, and 48 hours) were used to divide patients into early and late ambulation groups for each cutoff, and statistical analysis was performed to determine differences in postoperative outcomes between the corresponding groups.

RESULTS: Multivariate analysis showed no difference between the early and late ambulation groups for the 12-hour and 48-hour cutoffs; however, ambulation before 24 hours was associated with a decreased rate of severe complications as well as fewer adverse events overall. Patients who ambulated within 24 hours had a 4.1% rate of severe complications and a 22.1% rate of experiencing some adverse event (complication, return to the emergency department, and/or readmission). In comparison, 11.8% of patients who ambulated later experienced a severe complication (P = 0.026), while 36.1% of patients experienced some adverse event (P = 0.011).

CONCLUSIONS: Ambulation within 24 hours after colorectal surgery is associated with improved postoperative outcomes, particularly a decreased rate of severe complications.

PMID:36445980 | DOI:10.1177/00031348221142590

Oper Dent. 2022 Nov 28. doi: 10.2341/21-105-L. Online ahead of print.

ABSTRACT

OBJECTIVE: To assess the marginal and internal misfit of endocrowns fabricated from a resin matrix ceramic (CS), a partially crystallized (EMC), and a fully crystallized (ILS) lithium disilicate glass-ceramic.

METHODS AND MATERIALS: Thirty human premolar teeth restored with endocrowns were investigated. Three CAD/CAM materials were used (n=10 per group): CS, EMC, and ILS. Two-dimensional (2D) analysis of marginal and internal misfit was performed on micro-computed tomography scans before and after adhesive bonding. Further, three-dimensional (3D) analysis was performed to determine the total internal volume discrepancy. Surface roughness of the fitting surfaces of endocrowns was characterized using optical profilometry and scanning electron microscopy.

RESULTS: Adhesive bonding did not significantly affect marginal or internal misfit (p≥0.093). Differences in marginal misfit among the experimental groups were not statistically significant (p≥0.221). However, differences in 2D internal misfit were statistically significant; the CS group exhibited the largest internal misfit (p=0.001), while no significant difference was found between other groups (p=0.123). The largest discrepancies were observed at the pulpal floor and cervical region of all investigated specimens. No statistically significant difference was found in 3D misfit between ILS and EMC groups (p=0.711); however both exhibited statistically lower 3D misfit values compared to the CS group (p≤0.037). ILS endocrowns exhibited the smoothest and most homogenous fitting surface profile (p<0.001). However, there was no significant correlation between 2D internal misfit and the surface roughness (p≥0.082).

CONCLUSIONS: The choice of CAD/CAM material may influence the fitting accuracy of endocrowns. The investigated lithium disilicate glass-ceramics conferred superior internal fit for endocrowns compared to resin matrix ceramic.

PMID:36445975 | DOI:10.2341/21-105-L

Oper Dent. 2022 Nov 28. doi: 10.2341/21-202-L. Online ahead of print.

ABSTRACT

This study investigated the effects of environmental pH on the flexural properties of ion-releasing restorative materials (IRMs), including giomer (Beautifil-Bulk Restorative – BB), alkasite (Cention N – CN), bioactive composite (Activa – AB) and resin-modified glass ionomer (Riva Light Cure -RV) restoratives. A bio-inert resin-based composite (Filtek Bulk-fill Posterior – FB) served as the control. Stainless steel molds were used to fabricate 40 beam-shaped specimens (12mm × 2mm × 2mm) for each material. The specimens were finished, measured, and randomly distributed into four groups (n=10) and immersed in aqueous solutions of pH 3.0, pH 5.0, pH 6.8, and pH 10.0 at 37°C for 28 days. Specimens were then subjected to a uniaxial three-point bending flexural test with a load cell of 5 KN and a fixed deformation rate of 0.5 mm/min until fracture occurred. Flexural modulus and strength were statistically analyzed using analysis of variance/Dunnet T3’s test (p=0.05). Mean flexural modulus varied from (2.40±0.41 to 9.65±1.21 GPa), while mean flexural strength ranged from (21.56±2.78 to 163.86±13.13 MPa). Significant differences in flexural properties were observed among the various pH values and materials. All materials immersed in artificial saliva (pH 6.8) presented the highest flexural properties, except AB. The flexural strength of AB was significantly better when exposed to acidic environments. FB had better flexural properties than IRMs after exposure to a range of environmental pH values.

PMID:36445974 | DOI:10.2341/21-202-L

PLoS One. 2022 Nov 29;17(11):e0278285. doi: 10.1371/journal.pone.0278285. eCollection 2022.

ABSTRACT

Routine use of the Disease Activity Score-28 (DAS28) to assess the disease activity in rheumatoid arthritis (RA) is limited due to its dependency on laboratory investigations and the complex calculations involved. In contrast, the clinical disease activity index (CDAI) is simple to calculate, which makes the “treat to target” strategy for the management of RA more practical. We aimed to assess the validity of CDAI compared to DAS28 in RA patients in Sri Lanka. A total of 103 newly diagnosed RA patients were recruited, and their disease activity was calculated using DAS 28 and CDAI during the first visit to the clinic (0 months) and re-assessed at 4 and 9 months of follow-up visits. The validity of the CDAI, compared to DAS 28, was evaluated. Patients had a female preponderance (6:1) and a short symptom duration (mean = 6.33 months). Internal consistency reliability of CDAI, as assessed by Cronbach’s α test, was 0.868. Convergent validity was assessed by correlation and Kappa statistics. Strong positive correlations were observed between CDAI and DAS 28 at the baseline (0 months), 4 and 9 months of evaluation (Spearman’s r = 0.935, 0.935, 0.910, respectively). Moderate-good inter-rater agreements between the DAS-28 and CDAI were observed (Weighted kappa of 0.660, 0.519, and 0.741 at 0, 4, and 9 months respectively). Discriminant validity, as assessed by ROC curves at 0, 4th, and 9th months of the evaluation, showed the area under the curve (AUC) of 0.958, 0.979, and 0.910, respectively. The suggested cut-off points for different CDAI disease activity categories according to ROC curves were ≤ 4 (Remission), > 4 to ≤ 6 (low), > 6 to ≤ 18 (moderate), > 18 (high). These findings indicate that the CDAI has good concordance with DAS 28 in assessing the disease activity in RA patients, in this study sample.

PMID:36445922 | DOI:10.1371/journal.pone.0278285

PLoS One. 2022 Nov 29;17(11):e0277955. doi: 10.1371/journal.pone.0277955. eCollection 2022.

ABSTRACT

BACKGROUND: Globally, at least 4.7 million people die from being overweight or obese. In Ethiopia, the level of overweight and obesity among women grew from 3% to 8%. However, as far as my literature searching, studies concerning the spatial variation of overweight/obesity and factors associated are not researched in Ethiopia using geospatial techniques. Therefore, this study aimed to explore the spatial variation of overweight/obesity and factor associated among reproductive age group women in Ethiopia using geospatial techniques.

MOTHED: A total weighted sample of 10,928 reproductive age women were included in the study. ArcGIS version10.7 was used to explore the spatial variation of overweight/obesity. Bernoulli based model was used to analyze the purely spatial cluster detection of overweight/obesity through SaTScan version 9.6.1 software. Ordinary Least Square analysis and geographically weighted regression analysis was employed to assess the association between an outcome variable and explanatory variables by using ArcGIS 10.7 software. P value of less than 0.05 was used to declare statically significant.

RESULT: The spatial distribution of overweight/obesity in Ethiopia was clustered. Statistically, a significant-high hot spot overweight/obesity was identified at Addis Ababa, harrari, Dire Dawa. SaTScan identified 66 primary spatial clusters (RR = 4.17, P < 0.001) located at Addis Ababa, southeast amhara, central part of oromia region and northern part of SNNP region. In geographically weighted regression, rich wealth index, women’s age (35-39 and 40-44 years), watching TV, internet use and not working were statistically significant that affecting spatial variation of overweight/obesity.

CONCLUSION: In Ethiopia, overweight/obesity varies across the region. Statistically, significant-high hot spots of overweight/obesity were detected in Addis Ababa, Harari, Dire Dawa, some parts of Amhara and afar region, most of the Oromia and Somalia region, and the South Nation Nationality and People region of Ethiopia. Therefore, the ministry of health and the Ethiopian public health institute, try to initiate policies and practices that could include providing funding for physical education as well as recreational centers in communities most in need. In addition, public and private mass media create awareness of healthy lifestyles is promoted by health education regarding increased physical activity and reduced sedentary behavior through various media platforms.

PMID:36445917 | DOI:10.1371/journal.pone.0277955

PLoS One. 2022 Nov 29;17(11):e0277263. doi: 10.1371/journal.pone.0277263. eCollection 2022.

ABSTRACT

BACKGROUND: Adolescent girls have a greater nutrient demand and their poor dietary intake is associated with micronutrient deficiencies and poor maternal outcomes. Having information on micronutrient intake inadequacy in adolescent girls is critical for promoting healthy behavior and breaking the cycle of intergenerational malnutrition. Thus, this study assessed overall micronutrient intake inadequacy and associated factors among school adolescent girls in Meshenti town of Bahir Dar City Administration, North West Ethiopia.

METHODS: A school-based cross-sectional study was conducted among 401 adolescent girls from February 7 to 23, 2020. A Simple random sampling technique was used to select study participants. A multiple-pass 24-hour dietary recall with portion size estimation method and recommended dietary allowance cut-off point were used to assess micronutrient intake inadequacy. Overall micronutrient intake inadequacy was measured using the mean adequacy ratio. Nutrient databases were developed by ESHA FOOD PROCESSOR version 8.1 software. Data were entered into Epi-data version 3.1 and exported to SPSS version 23 for analysis. Multivariable logistic regression was performed to identify determinants of overall micronutrient intake inadequacy and an adjusted odds ratio at a p-value of less than 0.05 was used to see the strength of statistical association.

RESULTS: The prevalence of overall micronutrient intake inadequacy was 44.4% (95% CI: 39.7%-49.6%). Early adolescent age (AOR: 2.75, 95% CI: 1.71-4.42), food-insecure household (1.74, 95%CI: 1.087-2.784), low dietary diversity score (AOR = 2.83, 95% CI: 1.35-5.92), and high peer pressure on eating and body concern (AOR = 1.853, 95% CI: 1.201-2.857) were significantly associated factors with overall micronutrient intake inadequacy.

CONCLUSION: Findings of this study revealed that micronutrient intake inadequacy among adolescent girls was a high public health problem in the study area. Therefore, attention should be given to adolescent girls of the study area, especially the ones in the early adolescent age. Interventions should also focus on nutrition-sensitive activities to address food insecurity, a less diversified diet, and the negative impact of peer influence.

PMID:36445906 | DOI:10.1371/journal.pone.0277263