Search

The investigation of demographic bias in facial analysis applications is a topic of growing interest with achievements in face recognition and gender classification. State-of-the-art convolutional neural networks (CNN) and traditional machine learning models for locating facial landmarks have reached overall performance levels close to human annotation. However, recent studies demonstrated that these models presented performance gaps when applied to different populations, characterizing bias led by demographic features. Nevertheless, few studies have addressed this problem in face alignment and facial landmarks localization methods. In this work, we propose a multi-level face alignment approach settled on CNN models to reduce performance gaps among different populations. We created facial subunit CNN models tied to a facial subunit detector at a higher level. The proposal seeks to improve bad results caused by facial impairment, guided by the following assumptions: facial unit landmarks localization does not require global texture, and combining different facial unit models can improve the final model's variability. We applied the models in a balanced dataset mixing healthy controls and individuals with neurological disorders: amyotrophic lateral sclerosis and post-stroke. With fewer samples for training, our approach significantly reduced face alignment performance differences among those groups as compared to a state-of-the-art CNN model. [ABSTRACT FROM AUTHOR]

Glaucoma is an ocular disease that causes damage to the optic nerve, inducing successive narrowing of the visual field in affected patients due to an increased intraocular pressure, which can lead patients to blindness in an advanced stage without clinical reversal. For several years, the use of deep learning with convolutional neural networks (CNNs) has been successfully put into practice for several years. However, building a deep learning network requires an amount of experiments to find the fittest parameters, best choice of layers and an amount of available data. Thus it is not always able to produce satisfactory results due to the amount of parameters that need to be configured to adapt the CNN architecture to the problem in question, in most cases, with small datasets. Based on this scenario, this paper proposes and analyzes a CNN architecture construction from scratch, based on Neuroevolution of Augmenting Topologies for diagnosing glaucoma from fundus images. The method was evaluated with RIM-ONE and the combination of five glaucoma datasets in which we highlight 0.961 and 0.943 of f1-score, respectively for each dataset. [ABSTRACT FROM AUTHOR]

At the end of 2019, the World Health Organization (WHO) reported pneumonia that started in Wuhan, China, as a global emergency problem. Researchers quickly advanced in research to try to understand this COVID-19 and sough solutions for the front-line professionals fighting this fatal disease. One of the tools to aid in the detection, diagnosis, treatment, and prevention of this disease is computed tomography (CT). CT images provide valuable information on how this new disease affects the lungs of patients. However, the analysis of these images is not trivial, especially when researchers are searching for quick solutions. Detecting and evaluating this disease can be tiring, time-consuming, and susceptible to errors. Thus, in this study, we aim to automatically segment infections caused by COVID19 and provide quantitative measures of these infections to specialists, thus serving as a support tool. We use a database of real clinical cases from Pedro Ernesto University Hospital of the State of Rio de Janeiro, Brazil. The method involves five steps: lung segmentation, segmentation and extraction of pulmonary vessels, infection segmentation, infection classification, and infection quantification. For the lung segmentation and infection segmentation tasks, we propose modifications to the traditional U-Net, including batch normalization, leaky ReLU, dropout, and residual block techniques, and name it as Residual U-Net. The proposed method yields an average Dice value of 77.1% and an average specificity of 99.76%. For quantification of infectious findings, the proposed method achieves results like that of specialists, and no measure presented a value of ρ < 0.05 in the paired t-test. The results demonstrate the potential of the proposed method as a tool to help medical professionals combat COVID-19. fight the COVID-19. [ABSTRACT FROM AUTHOR]

In-vitro fertilization is a routine livestock-breeding technique widely used around the world. Several studies have reported the interaction of bovine viral-diarrhea virus (BVDV) with gametes and in-vitro-produced (IVP) bovine embryos. Since, gene expression in BVDV-infected IVP bovine embryos is scarcely addressed. The aim of this work was to evaluate the differential expression of genes involved in immune and inflammatory response. Groups of 20–25 embryos on Day 6 (morula stage) were exposed (infected) or not (control) to an NCP-BVDV strain in SOF medium. After 24 h, embryos that reached expanded blastocyst stage were washed. Total RNA of each embryo group was extracted to determine the transcription levels of 9 specific transcripts related with antiviral and inflammatory response by SYBR Green real time quantitative (RT-qPCR). Culture media and an aliquot of the last embryos wash on Day 7 were analyzed by titration and virus isolation, respectively. A conventional PCR confirmed BVDV presence in IVP embryos. A significantly higher expression of interferon-α was observed in blastocysts exposed to NCP-BVDV compared to the controls (p < 0.05). In this study, the upregulation of INFα and TLR7 genes involved in inflammatory and immune response in BVDV-infected IVP bovine embryos is a new finding in this field. This differential expression suggest that embryonic cells could function in a manner like immune cells by recognizing and responding early to interaction with viral pathogens. These results provide new insights into the action of BVDV on the complex molecular pathways controlling bovine early embryonic development. [ABSTRACT FROM AUTHOR]

Automatic and reliable prostate segmentation is an essential prerequisite for assisting the diagnosis and treatment, such as guiding biopsy procedure and radiation therapy. Nonetheless, automatic segmentation is challenging due to the lack of clear prostate boundaries owing to the similar appearance of prostate and surrounding tissues and the wide variation in size and shape among different patients ascribed to pathological changes or different resolutions of images. In this regard, the state-of-the-art includes methods based on a probabilistic atlas, active contour models, and deep learning techniques. However, these techniques have limitations that need to be addressed, such as MRI scans with the same spatial resolution, initialization of the prostate region with well-defined contours and a set of hyperparameters of deep learning techniques determined manually, respectively. Therefore, this paper proposes an automatic and novel coarse-to-fine segmentation method for prostate 3D MRI scans. The coarse segmentation step combines local texture and spatial information using the Intrinsic Manifold Simple Linear Iterative Clustering algorithm and probabilistic atlas in a deep convolutional neural networks model jointly with the particle swarm optimization algorithm to classify prostate and non-prostate tissues. Then, the fine segmentation uses the 3D Chan-Vese active contour model to obtain the final prostate surface. The proposed method has been evaluated on the Prostate 3T and PROMISE12 databases presenting a dice similarity coefficient of 84.86%, relative volume difference of 14.53%, sensitivity of 90.73%, specificity of 99.46%, and accuracy of 99.11%. Experimental results demonstrate the high performance potential of the proposed method compared to those previously published. [ABSTRACT FROM AUTHOR]

The aim of the present study was to characterize the specific immune response in prepubertal female calves inoculated with Neospora caninum. Forty-eight N. caninum-seronegative 6-month-old Angus female calves were randomly allocated into two groups: group A calves were inoculated subcutaneously (sc) with 1 × 106 tachyzoites of the low virulence NC-Argentina LP1 isolate in sterile phosphate-buffered saline (PBS); group B calves were mock inoculated sc with sterile PBS. Calves from group A developed a specific immune response characterized by the production of IgG antibodies and the expression of IFN-γ and TNF-α cytokines. Animals did not present any febrile reaction or reactions at the site of inoculation. Although chronic N. caninum infection was developed in 50% of calves of group A after inoculation, according to the presence of antibodies against rNc-SAG4, antigen characteristic of bradyzoites, N. caninum antibodies dropped below the cut-off of ELISA from day 210 post-inoculation onwards. Future trials using the same group of inoculated animals will allow the characterization of the evolution of the immune response during pregnancy and to determine whether the immunization with the local isolate is able to prevent congenital transmission and to protect against heterologous challenges. [ABSTRACT FROM AUTHOR]

Breast cancer is a global health problem which mainly affects the female population. It is known that early detection increases the chances of effective treatment, improving the disease prognosis. It remains a challenge to detect the lesion with high detection rate and ensure, at the same time, low rates of false positives. Aiming this objective, this work proposes an efficient method for detection of mass regions on digitized mammograms though diversity analysis, geostatistical and concave geometry (Alpha Shapes). We evaluate the detection rate for each feature extraction using Support Vector Machine in MIAS and DDSM database, with 74 and 621 mammograms, respectively, all containing at least one mass region. The obtained results are promising, reaching 97.30% of detection rate and 0.89 false positive per image for MIAS database and also 91.63% of detection rate and 0.86 false positive per image for DDSM database. Specifically, in DDSM obtaining high detection rate and low rate of false positives when using concave geometry to extract features in a large database. [ABSTRACT FROM AUTHOR]