RESUMEN
Facial expression recognition(FER) is a hot topic in computer vision, especially as deep learning based methods are gaining traction in this field. However, traditional convolutional neural networks (CNN) ignore the relative position relationship of key facial features (mouth, eyebrows, eyes, etc.) due to changes of facial expressions in real-world environments such as rotation, displacement or partial occlusion. In addition, most of the works in the literature do not take visual tempos into account when recognizing facial expressions that possess higher similarities. To address these issues, we propose a visual tempos 3D-CapsNet framework(VT-3DCapsNet). First, we propose 3D-CapsNet model for emotion recognition, in which we introduced improved 3D-ResNet architecture that integrated with AU-perceived attention module to enhance the ability of feature representation of capsule network, through expressing deeper hierarchical spatiotemporal features and extracting latent information (position, size, orientation) in key facial areas. Furthermore, we propose the temporal pyramid network(TPN)-based expression recognition module(TPN-ERM), which can learn high-level facial motion features from video frames to model differences in visual tempos, further improving the recognition accuracy of 3D-CapsNet. Extensive experiments are conducted on extended Kohn-Kanada (CK+) database and Acted Facial Expression in Wild (AFEW) database. The results demonstrate competitive performance of our approach compared with other state-of-the-art methods.
Asunto(s)
Expresión Facial , Redes Neurales de la Computación , Humanos , Grabación en Video/métodos , Reconocimiento Facial Automatizado/métodos , Aprendizaje Profundo , Emociones/fisiología , Reconocimiento Facial/fisiología , Imagenología Tridimensional/métodosRESUMEN
Both thermoelectric and mechanical properties are important to the practical applications of thermoelectric materials. Herein, we develop a strategy for alloying KCu7S4 to improve the dimensionless figure of merit (zT), compressive strength, and Vickers hardness of polycrystalline SnSe. Through chemical synthesis and particle mixing in solutions, powders with SnSe nanoparticles and KCu7S4 nanowires are produced, and the subsequent spark plasma sintering triggers the reaction between the two chalcogenides, resulting in the formation of Cu2SnSe3 nanoparticles and substitution of Cu and S in the SnSe matrix. The composition tuning and secondary phase formation effectively enhance the power factor and diminish the lattice thermal conductivity, leading to a maximum zT of 1.13 at 823 K for the optimal sample, which is improved by 135% over that of SnSe. Simultaneously, the compressive strength and hardness are also enhanced, as exemplified by a high compressive strength of 135 MPa that is enhanced by â¼81% compared to that of SnSe. The current study demonstrates effective composite and composition design toward enhanced thermoelectric and mechanical performance in polycrystalline SnSe.
RESUMEN
Silent speech decoding (SSD), based on articulatory neuromuscular activities, has become a prevalent task of brain-computer interfaces (BCIs) in recent years. Many works have been devoted to decoding surface electromyography (sEMG) from articulatory neuromuscular activities. However, restoring silent speech in tonal languages such as Mandarin Chinese is still difficult. This paper proposes an optimized sequence-to-sequence (Seq2Seq) approach to synthesize voice from the sEMG-based silent speech. We extract duration information to regulate the sEMG-based silent speech using the audio length. Then, we provide a deep-learning model with an encoder-decoder structure and a state-of-the-art vocoder to generate the audio waveform. Experiments based on six Mandarin Chinese speakers demonstrate that the proposed model can successfully decode silent speech in Mandarin Chinese and achieve a character error rate (CER) of 6.41% on average with human evaluation.
RESUMEN
Taste is a key sense that helps identify different food types and most of this work was carried out on primary tastes rather than generating different flavors. In this work, we proposed a plan to create other flavors rather than primary tastes, adjusted the electrical (40-180 µA) and thermal stimulation (20-38 °C and 38-20 °C), and revealed the digital coding for multi-flavors. Our results showed that different combinations of digital coding could generate different flavors and that tastes related to different stimuli are easy to develop. The novelty of this work is to design other types of flavors and primary tastes. The experimental results demonstrated that the novel method proposed for digital taste coding could realize primary tastes (sweet, sour, salty, spicy, and mint) and mixed flavors. Furthermore, some innovative sensations have been realized, which are sprite, soda water, sweet-sour, salty-sweet, and salty-mint sensations. We presume that this innovation could digitally enhance various flavors.
Asunto(s)
Gusto , Lengua , Gusto/fisiologíaRESUMEN
Recently, research on the energy harvesting floor is attracting more and more attention due to its possible application in the smart house, invasion monitoring, internet of things, etc. This paper introduced a design and comparative study of a small-stroke piezoelectric energy harvesting floor based on a multi-layer piezoelectric beam structure. The multi-layer piezoelectric beams are designed based on simply supported beams in an interdigitated manner. Theoretical analysis is explored to find out the beam number and layer number of the structure. Through this design, the input power from the human footsteps was effectively utilized and transformed into electrical power. The designed piezoelectric energy harvesting floor structure was tested by our designed stepping machine, which can simulate the stepping effect of a walking human on the floor with different parameters such as stepping frequency. Comparative studies of the energy harvester are carried out regarding different stepping frequencies, external circuits, and initial beam shapes. The experimental results showed that the maximum output power of a group of four-layer prototypes was 960.9 µW at a stroke of 4 mm and a step frequency of 0.83 Hz, with the beams connected in parallel.
RESUMEN
Rotational energy harvesting technology has attracted more and more attention recently. This paper presents a piezoelectric rotational energy harvester that can be mounted with an offset distance from the rotation center. The piezoelectric energy harvester is designed to be dynamically excited by the force due to gravity, which causes the piezoelectric cantilever beams in the harvester to vibrate periodically as the harvester rotates. A novel design of the harvester structure with a hollow mass is proposed and analyzed in this paper. Experiments were performed to investigate the design and analysis. A power output of 106~2308 µW can be achieved at the rotating frequencies of 0.79~14 Hz with a piezoelectric cantilever beam in the prototyped energy harvester. Results showed that the prototyped harvester can be mounted on a rotating wheel hub and output sufficient power in a wide frequency range for wireless monitoring sensors.
RESUMEN
Bornite (Cu5 FeS4 ) is an Earth-abundant, nontoxic thermoelectric material. Herein, twin engineering and Se alloying are combined in order to further improve its thermoelectric performance. Cu5 FeS4-x Sex (0 ≤ x ≤ 0.4) icosahedral nanoparticles, containing high-density twin boundaries, have been synthesized by a colloidal method. Spark plasma sintering retains twin boundaries in the pellets sintered from Cu5 FeS4-x Sex colloidal powders. Thermoelectric property measurement demonstrates that alloying Se increases the carrier concentration, leading to much-improved power factor in Se-substituted Cu5 FeS4 , for example, 0.84 mW m-1 K-2 at 726 K for Cu5 FeS3.6 Se0.4 ; low lattice thermal conductivity is also achieved, due to intrinsic structural complexity, distorted crystal structure, and existing twin boundaries and point defects. As a result, a maximum zT of 0.75 is attained for Cu5 FeS3.6 Se0.4 at 726 K, which is about 23% higher than that of Cu5 FeS4 and compares favorably to that of reported Cu5 FeS4 -based materials. In addition, the Cu5 FeS4-x Sex samples containing twin boundaries also obtain improved hardness compared to the ones fabricated by melting-annealing or ball milling. This work demonstrates an effective twin engineering-composition tuning strategy toward enhanced thermoelectric and mechanical properties of Cu5 FeS4 -based materials.
RESUMEN
A novel approach to quantitatively recognize the intensity of primary taste stimuli was explored based on surface electromyography (sEMG). We captured sEMG samples under stimuli of primary taste with different intensities and quantitatively recognized preprocessed samples with Support Vector Machine (SVM). The feasibility of quantitatively recognizing the intensity of Sour, Bitter, and Salty was verified. The sEMG signals were acquired under the stimuli of citric acid (aq), sucrose (aq), magnesium chloride (aq), sodium chloride (aq), and sodium glutamate (aq) with different concentrations, for five types of primary tastes: Sour, Sweet, Bitter, Salty, and Umami, whose order was fixed in this article. The acquired signals were processed with a method called Quadratic Variation Reduction to remove baseline wandering, and an adaptive notch to remove power frequency interference. After extracting 330 features for each sample, an SVM regressor with five-fold cross-validation was performed and the model reached R2 scores of 0.7277, 0.1963, 0.7450, 0.7642, and 0.5055 for five types of primary tastes, respectively, which manifested the feasibilities of the quantitative recognitions of Sour, Bitter, and Salty. To explore the facial responses to taste stimuli, we summarized and compared the muscle activities under stimuli of different taste types and taste intensities. To further simplify the model, we explored the impact of feature dimensionalities and optimized the feature combination for each taste in a channel-wise manner, and the feature dimensionality was reduced from 330 to 210, 120, 210, 260, 170 for five types of primary tastes, respectively. Lastly, we analyzed the model performance on multiple subjects and the relation between the model's performance and the number of experiment subjects. This study can provide references for further research and applications on taste stimuli recognition with sEMG.
Asunto(s)
Sacarosa , Gusto , Electromiografía , Humanos , Cloruro de Sodio , Glutamato de SodioRESUMEN
Cu2S, consisting of Earth-abundant and eco-friendly elements, is a promising candidate for thermoelectric energy conversion. The stoichiometric Cu2S has intrinsically low thermal conductivity together with meagre electrical conductivity. Herein, in order to improve the electrical properties, we develop a facile solvothermal method for synthesizing a series of Cu2-xS with tunable phase compositions. With reduction in the CuCl/Na2S molar ratio in the solution precursors from 2:1 to 1.75:1, the solvothermal products transform from single-phase Cu2S into Cu2S-Cu1.96S composites, which are preserved in the hot-pressed pellets. The decreased CuCl/Na2S molar ratio results in increased room-temperature Hall carrier concentration, leading to enhanced electrical conductivity. Specifically, the pellet corresponding to the 1.75:1 CuCl/Na2S precursor ratio obtains a power factor of 8.04 µW cm-1 K-2 at 815 K, which is â¼735% of the value obtained by stoichiometric Cu2S. Along with the moderate thermal conductivity, this sample achieves an enhanced zT value of 0.87 at 815 K, representing a â¼250% increase when compared to stoichiometric Cu2S. This study demonstrates an effective strategy in enhancing the thermoelectric performance of solution-synthesized Cu2S-based materials by phase tuning.
RESUMEN
Based on surface electromyography (sEMG), a novel recognition method to distinguish six types of human primary taste sensations was developed, and the recognition accuracy was 74.46%. The sEMG signals were acquired under the stimuli of no taste substance, distilled vinegar, white granulated sugar, instant coffee powder, refined salt, and Ajinomoto. Then, signals were preprocessed with the following steps: sample augments, removal of trend items, high-pass filter, and adaptive power frequency notch. Signals were classified with random forest and the classifier gave a five-fold cross-validation accuracy of 74.46%, which manifested the feasibility of the recognition task. To further improve the model performance, we explored the impact of feature dimension, electrode distribution, and subject diversity. Accordingly, we provided an optimized feature combination that reduced the number of feature types from 21 to 4, a preferable selection of electrode positions that reduced the number of channels from 6 to 4, and an analysis of the relation between subject diversity and model performance. This study provides guidance for further research on taste sensation recognition with sEMG.
Asunto(s)
Algoritmos , Gusto , Electrodos , Electromiografía , HumanosRESUMEN
SnTe is proposed to be an intriguing low-toxicity alternative to PbTe. Herein, we report the diminished lattice thermal conductivity (κL) and enhanced zT of SnTe by way of vacancy engineering. (SnTe)1-x(Sb2Te3)x (x = 0.03, 0.06, and 0.10) and (SnTe)1-y(Sb2Se3)y (y = 0.03 and 0.06) were synthesized by blending and sintering their solution-synthesized nano/microstructures (i.e., SnTe octahedral particles, Sb2Te3 nanoplates, and Sb2Se3 nanorods). Benefiting from the chemical reactions during sintering, single-phase SnTe-based solid solutions were formed when x or y is not higher than 0.06, into which tunable concentrations of Sn vacancies were introduced. Such vacancies significantly enhance phonon scattering, leading to the sharply reduced room temperature κL of 1.40 and 1.26 W m-1 K-1 for x = 0.06 and y = 0.06 samples, respectively, as compared to 3.73 W m-1 K-1 for pristine SnTe. Enabled by point defects with the highest concentration and SnSb2Te4 secondary phase, (SnTe)0.90(Sb2Te3)0.10 sample obtains the lowest κL of 0.70 W m-1 K-1 at 813 K. Ultimately, maximum zT values of 0.6 and 0.7 at 813 K are achieved in (SnTe)0.90(Sb2Te3)0.10 and (SnTe)0.94(Sb2Se3)0.06, respectively. This study demonstrates the effectiveness of vacancy engineering in improving zT of SnTe-based materials.
RESUMEN
BACKGROUND: Fibronectin type III domain containing 3B (FNDC3B) acts as an oncogene in various cancers, and abnormal expression of FNDC3B has been found in colorectal cancer (CRC). Our study aimed to illustrate the role of FNDC3B in CRC development. METHODS: Through RT-qPCR and western blotting assays, the mRNA and protein expressions of target genes were measured. CCK-8 and MTT methods were used to detect cell proliferation. Invasion ability was determined using Transwell assay. TargetScan platform and luciferase reporter gene assay were performed to predict and validate the bindings between FNDC3B and miR-125a-5p or miR-217. Besides, the expression correlation was measured by Pearson's Correlation analysis. RESULTS: We found that FNDC3B was significantly upregulated in CRC tissues and tumor cell lines, and high expression of FNDC3B predicted a poor survival outcome. The bindings between FNDC3B and miR-125a-5p and miR-217 were respectively at the motifs of CUCAGGG and AUGCAGU. MiR-125a-5p and miR-217 were downregulated in CRC tissues, and both were negatively correlated with FNDC3B expression. Subsequently, the downregulated miR-125a-5p and miR-217 were confirmed as contributors FNDC3B upregulation in CRC. A loss-of-function assay demonstrated that FNDC3B knockdown inhibited the proliferation of CRC cells, while FNDC3B overexpression promoted the proliferation and invasion of tumor cells. Besides, we validated that PI3K/mTOR signaling was involved in the regulation of FNDC3B on the proliferation and invasion of CRC cells. CONCLUSION: Generally, our findings demonstrated that FNDC3B facilitated cell proliferation and invasion via PI3K/mTOR signaling, and further promoted CRC progression. The novel miR-125a-5p/FNDC3B and miR-217/FNDC3B axes might be new targets for CRC prognosis and therapy.
RESUMEN
Proline-rich protein 11 (PRR11) has been shown to play an critical roles in the development of cancer. However, the clinical significance and the biological role of PRR11 in hepatocellular carcinoma (HCC) remains unknown. The present study aimed to investigate the expression pattern, prognostic value and the biological role of PRR11 in HCC. PRR11 expression in 80 HCC surgical specimens was examined, and its clinical significance was analyzed. The role of PRR11 in cell proliferation, colony formation, migration and invasion were also determined. The results showed that PRR11 mRNA was significantly up-regulated in 56.25% (45/80) HCC from that in matched adjacent non-tumor tissues. High PRR11 was correlated with tumor size (Pâ¯=â¯0.01) and TNM stage (Pâ¯=â¯0.006). Patients with higher PRR11 expression had poor overall survival time (Pâ¯<â¯0.001). Furthermore, PRR11 silencing obviously inhibited cell proliferation, colony formation, as well as cell migration and invasion of HCC cell lines in vitro. Mechanistically, knockdown of PRR11 significantly decreased the expression of ß-catenin, cyclinD1, c-myc and N-cadherin in HCC cell lines. Additionally, the inhibitory effects of PRR11 silencing on cell migration was significantly enhanced by ß-catenin inhibition. PRRl1 mRNA expression was found positively correlated with ß-catenin (Râ¯=â¯0.5472, Pâ¯Ëâ¯0.0001), c-myc (Râ¯=â¯0.5527, Pâ¯Ëâ¯0.0001) and cyclinD1 (Râ¯=â¯0.3948, Pâ¯=â¯0.0003) in HCC tissues. Collectively, our data demonstrate that PRR11 plays an oncogenic role in HCC progression, through activating the Wnt/ß-catenin signaling pathway, and may represent a valuable prognostic marker and therapeutic target for HCC.