74227 research outputs found
Sort by
A multi-task framework for breast cancer segmentation and classification in ultrasound imaging
Get PDFBackground: Ultrasound (US) is a medical imaging modality that plays a crucial role in the early detection
of breast cancer. The emergence of numerous deep learning systems has offered promising avenues for the
segmentation and classification of breast cancer tumors in US images. However, challenges such as the absence
of data standardization, the exclusion of non-tumor images during training, and the narrow view of single-task
methodologies have hindered the practical applicability of these systems, often resulting in biased outcomes.
This study aims to explore the potential of multi-task systems in enhancing the detection of breast cancer
lesions.
Methods: To address these limitations, our research introduces an end-to-end multi-task framework designed
to leverage the inherent correlations between breast cancer lesion classification and segmentation tasks.
Additionally, a comprehensive analysis of a widely utilized public breast cancer ultrasound dataset named
BUSI was carried out, identifying its irregularities and devising an algorithm tailored for detecting duplicated
images in it.
Results: Experiments are conducted utilizing the curated dataset to minimize potential biases in outcomes.
Our multi-task framework exhibits superior performance in breast cancer respecting single-task approaches,
achieving improvements close to 15% in segmentation and classification. Moreover, a comparative analysis
against the state-of-the-art reveals statistically significant enhancements across both tasks.
Conclusion: The experimental findings underscore the efficacy of multi-task techniques, showcasing better
generalization capabilities when considering all image types: benign, malignant, and non-tumor images.
Consequently, our methodology represents an advance towards more general architectures with real clinical
applications in the breast cancer fiel
Exosomes in ocular health: recent insights into pathology, diagnostic applications and therapeutic functions
No full textInstituto de Desarrollo Economico del Principado de Asturias (IDEPA), Government of the Principado de Asturias (Spain) [IDE/2022/000641]; Spanish Ministry of Universities [FPU20/06016
Patch antennas characterization for enhanced microwave imaging
Get PDFThis dataset corresponds to the measurements of two microstrip patch antennas, collected using the facility described in [1]. The available measurements contained within the dataset allow a complete characterization of the field radiated by these antennas. These fields can be introduced in enhanced microwave imaging algorithms that consider the field radiated by the transmitting and receiving antennas of the microwave imaging system [2] (modified Delay and Sum algorithm), [3] (modified Phase Shift Migration imaging algorithm). The patch antennas that are characterized are the ones presented in [4].
The files "Emeas_leftPatch_z30cm.zip" and "Emeas_rightPatch_z30cm.zip", correspond to the copolar component (Ex component) of the electric field radiated by the patch antennas (left and right as depicted in the figure “PictureDataset.png”). More precisely, it corresponds to the S21 parameter, which is proportional to the electric field radiated by the patch antenna.
Measurements were conducted within the frequency range from 22 GHz to 28 GHz, with a frequency step of 15 MHz. Measurements were collected on a domain of size Lx × Ly = 70 cm × 70 cm, discretized every δx,y = 5 mm (0.42 wavelengths at the center frequency of 25 GHz). The distance between the patch antenna (antenna under test) and the measurement plane was 30 cm. An Open-Ended Waveguide was used as probe antenna.
The measured electric field was backpropagated from the measurement plane to the patch antenna aperture plane using the backpropagation algorithm described in [5]. The electric fields on the aperture plane are provided in the files “Eap_leftPatch.zip” and “Eap_rightPatch.zip” for the left and right patch antennas, respectively. The aperture fields were calculated on a planar domain having the same size as the measurement plane, that is, Lx × Ly = 70 cm × 70 cm, and also discretized every δx,y = 5 mm.
[1] A. Arboleya, Y. Alvarez, and F. Las-Heras, “Millimeter and submillimeter planar measurement setup,” in 2013 IEEE Antennas and Propagation Society International Symposium (APSURSI), 2013, pp. 1–2.
[2] Y. Alvarez Lopez and F. Las-Heras, “On the use of an equivalent currents-based technique to improve electromagnetic imaging,” IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 8004113, 2022.
[3] Y. Alvarez López and F. Las-Heras Andrés, "Improved Methods for Fourier-based Microwave Imaging," Sensors, Vol. 23, pp. 9250, 2023.
[4] A. F. Berdasco, J. Laviada, M. E. de Cos Gómez and F. Las-Heras, “Performance Evaluation of Millimeter-Wave Wearable Antennas for Electronic Travel Aid,” in IEEE Transactions on Instrumentation and Measurement, vol. 72, pp. 1-10, 2023, Art no. 4507510, doi: 10.1109/TIM.2023.3320736.
[5] J. Hanfling, G. Borgiotti, and L. Kaplan, “The backward transform of the near field for reconstruction of aperture fields,” in 1979 Antennas and Prop. Society Intl. Symposium, vol. 17, 1979, pp. 764–767.Ministerio de Ciencia e Innovación of Spain, Agencia Estatal de Investigación of Spain, and Fondo Europeo de Desarrollo Regional (FEDER). Grant Number: PID2021-122697OB-I00 (“META-IMAGER”
Navigational object-location memory assessment in real and virtual environments: a systematic review
Get PDFNavigational object-location memory (OLM) is a form of spatial memory involving actual or virtual body
displacement for repositioning previously encoded objects within an environment. Despite its potential for
higher ecological validity measures, navigational OLM has been less frequently assessed than static OLM. The
present systematic review aims to characterize the methodology and devices used for OLM assessment in
navigational real and virtual environments and synthesize recent literature to offer a comprehensive overview of
OLM performance in both pathological and non-pathological adult samples. A search through four different
databases was conducted, identifying 39 studies. Most studies assessed navigational OLM in healthy adults by 2-
dimensional or 3-dimensional computerized tasks, although immersive Virtual Reality (VR) devices were also
frequently employed. Small environments and objects with high-semantic value were predominantly used, with
assessment mainly conducted immediately after learning through free-recall tasks. The findings revealed that
healthy samples outperformed clinical ones in navigational OLM. Men showed superior performance compared
to women when cues or landmarks were used, but this advantage disappeared in their absence. Better results
were also noted with shorter intervals between learning and recall. Fewer OLM errors occurred in real environments compared to both immersive and non-immersive VR. Influences of environmental features, object
semantics, and participant characteristics on OLM performance were also observed. These results highlight the
need for standardized methodologies, the inclusion of a broader age range in populations, and careful control
over the devices, environments, and objects used in navigational OLM assessments.Gobierno de Aragon (Departamento de Ciencia, Universidad y Sociedad del Conocimiento) for group S31_23R; Conselleria d'Innovacio, Universitats, Ciencia i Societat Digital de la Generalitat Valenciana [GVA-COVID19/2021/025]. The contribution of TL has been supported by the Programa Severo Ochoa de la Consejeria de Ciencia, Empresas, Formacion y Empleo (ref. PA-23-BP22\u2013005
Endangered, exploited glass eels (anguilla anguilla) with critical levels of heavy metals and microplastics reveal both shipping and plastic spill threats
No full textA Dynamic study of the Single Active Bridge Converter
Get PDFIn this paper, a thorough study of the dynamic behavior of the Single Active Bridge (SAB) converter is presented. The SAB converter can be considered a unidirectional version of the Dual Active Bridge (DAB) converter, which has been extensively analyzed in recent years. However, the dynamics of the SAB converter differs from the one corresponding to the DAB converter and has not been addressed so far. The SAB converter can operate in two different conduction modes, namely Discontinuous Conduction Mode (DCM) and Continuous Conduction Mode (CCM). The SAB operating in DCM presents the same static and dynamic behavior as the Phase-Shifted Controlled Full Bridge (PSFB) converter if the value of the inductor of the SAB is the same as the value of the output inductor of the PSFB referred to the transformer primary side. However, the dynamic behavior of the SAB converter in CCM is different and must be analyzed in detail. As a result of the analysis carried out, average small-signal linear models have been obtained for both conduction modes. These models are compared with the ones obtained for other similar converters and they have been validated with simulation and experimental results.This work has been supported by the European Union under project UE-23-POWERIZED-101096387, the Spanish Government under project MCINN-22-TED2021-130939B and by the Principality of Asturias grant “Severo Ochoa” BP21-114
EOG compression in polysomnographic recordings based on the Lempel-Ziv-Welch algorithm
Get PDFNocturnal polysomnography (PSG) is a neurophysiological technique that studies sleep by recording multiple
physiological parameters. One is the electrical signal, called the electrooculogram (EOG), generated from eye
movement. An extensive PSG signal recording, typically around 8 h, requires a massive volume of data to be
transmitted and stored; compression is therefore required. This study aims to compress EOG signals effectively,
providing high-quality reconstruction with low bit rates and acceptable distortions. The Sleep Disorders Research
Center dataset is employed to verify the applicability of the devised method. The solution is founded on the
Lempel-Ziv-Welch (LZW) algorithm, developed with MATLAB software. The signal is compressed using this algorithm and subsequently reconstructed. The algorithm’s performance is evaluated using five parameters:
compression performance (CP), L2 energy retained in the compressed signal, percent root-mean-square difference
(PRD) in the reconstruction, compressed signal size, and runtime. The findings of the experiment, which used 22
EOGs of different subjects, comprising 11 individuals with psychophysiological insomnia and 11 individuals
without this condition, demonstrated that the LZW algorithm produced an average CP of 84.65% while retaining
almost 100.44% of the signal energy and a PRD of 6.20% in the reconstructed signal
Does pregabalin offer potential as a first-line therapy for generalized anxiety disorder? A meta-analysis of efficacy, safety, and cost-effectiveness
No full textRol de la psicología como acompañamiento en tratamientos de reproducción asistida: un diseño de intervención basado en ACT
Get PDFLa infertilidad representa una crisis emocional, psicológica y médica significativa para quienes la padecen. Este impacto se ve intensificado en el contexto de los tratamientos de reproducción asistida (RA), donde la incertidumbre y la falta de control pueden afectar al bienestar de las mujeres y sus parejas. En respuesta a esta necesidad, se presenta un diseño de intervención psicológica como acompañamiento, basado en la Terapia de Aceptación y Compromiso (ACT), como enfoque innovador y alternativo a los modelos actuales. El programa consta de seis sesiones estructuradas que integra estrategias de aceptación del malestar, mindfulness, clarificación de valores personales y defusión cognitiva. Mediante esta intervención, se busca humanizar los tratamientos de RA, sentando las bases para su futura implementación contribuyendo al bienestar y calidad de vida a través de la prevención primaria
