RESUMO
Sparse Coding (SC) has been widely studied and shown its superiority in the fields of signal processing, statistics, and machine learning. However, due to the high computational cost of the optimization algorithms required to compute the sparse feature, the applicability of SC to real-time object recognition tasks is limited. Many deep neural networks have been constructed to low fast estimate the sparse feature with the help of a large number of training samples, which is not suitable for small-scale datasets. Therefore, this work presents a simple and efficient fast approximation method for SC, in which a special single-hidden-layer neural network (SLNNs) is constructed to perform the approximation task, and the optimal sparse features of training samples exactly computed by sparse coding algorithm are used as ground truth to train the SLNNs. After training, the proposed SLNNs can quickly estimate sparse features for testing samples. Ten benchmark data sets taken from UCI databases and two face image datasets are used for experiment, and the low root mean square error (RMSE) results between the approximated sparse features and the optimal ones have verified the approximation performance of this proposed method. Furthermore, the recognition results demonstrate that the proposed method can effectively reduce the computational time of testing process while maintaining the recognition performance, and outperforms several state-of-the-art fast approximation sparse coding methods, as well as the exact sparse coding algorithms.
RESUMO
Brain functional connectivity has been routinely explored to reveal the functional interaction dynamics between the brain regions. However, conventional functional connectivity measures rely on deterministic models fixed for all participants, usually demanding application-specific empirical analysis, while deep learning approaches focus on finding discriminative features for state classification, thus having limited capability to capture the interpretable functional connectivity characteristics. To address the challenges, this study proposes a self-supervised triplet network with depth-wise attention (TripletNet-DA) to generate the functional connectivity: 1) TripletNet-DA firstly utilizes channel-wise transformations for temporal data augmentation, where the correlated & uncorrelated sample pairs are constructed for self-supervised training, 2) Channel encoder is designed with a convolution network to extract the deep features, while similarity estimator is employed to generate the similarity pairs and the functional connectivity representations with prominent patterns emphasized via depth-wise attention mechanism, 3) TripletNet-DA applies Triplet loss with anchor-negative similarity penalty for model training, where the similarities of uncorrelated sample pairs are minimized to enhance model's learning capability. Experimental results on pathological EEG datasets (Autism Spectrum Disorder, Major Depressive Disorder) indicate that 1) TripletNet-DA demonstrates superiority in both ASD discrimination and MDD classification than the state-of-the-art counterparts in various frequency bands, where the connectivity features in beta & gamma bands have respectively achieved the accuracy of 97.05%,98.32% for ASD discrimination, 89.88%,91.80% for MDD classification in the eyes-closed condition, and 90.90%,92.26% for MDD classification in the eyes-open condition, 2) TripletNet-DA enables to uncover significant differences of functional connectivity between the ASD EEG and the TD ones, and the prominent connectivity links are in accordance with the empirical findings that frontal lobe demonstrates more connectivity links and significant frontal-temporal connectivity occurs in the beta band, thus providing potential biomarkers for clinical ASD analysis.
RESUMO
HD map reconstruction is crucial for autonomous driving. LiDAR-based methods are limited due to expensive sensors and time-consuming computation. Camera-based methods usually need to perform road segmentation and view transformation separately, which often causes distortion and missing content. To push the limits of the technology, we present a novel framework that reconstructs a local map formed by road layout and vehicle occupancy in the bird's-eye view given a front-view monocular image only. We propose a front-to-top view projection (FTVP) module, which takes the constraint of cycle consistency between views into account and makes full use of their correlation to strengthen the view transformation and scene understanding. In addition, we apply multi-scale FTVP modules to propagate the rich spatial information of low-level features to mitigate spatial deviation of the predicted object location. Experiments on public benchmarks show that our method achieves various tasks on road layout estimation, vehicle occupancy estimation, and multi-class semantic estimation, at a performance level comparable to the state-of-the-arts, while maintaining superior efficiency.
RESUMO
A nasopharyngeal carcinoma (NPC) mass screening trial using a combination of immunoglobulin A antibodies to Epstein-Barr virus capsid antigen and nuclear antigen-1 by enzyme-linked immunosorbent assay in addition to indirect mirror examination in the nasopharynx and/or lymphatic palpation (IMLP) was conducted in southern China. Cantonese aged 30-59 years residing in 2 cities randomly selected by cluster sampling, Sihui and Zhongshan, were invited to participate in this screening from May 2008 through May 2010. Participants were offered fiberoptic endoscopy examination and/or pathologic biopsy if their serologic tests reached our predefined level of high risk or if results from the physical examination indicated possible cancer (i.e., were IMLP positive). A total of 28,688 individuals were voluntarily screened in the initial round. The overall NPC detection rate was 0.14% (41/28,688) with an early diagnosis rate of 68.3% (28/41) during the first year of follow-up. Thirty-eight of 41 cases (92.7%) were detected among the high-risk group, and 7 of 41 cases (17.1%) were detected among the IMLP-positive group. The 2 Epstein-Barr virus serologic tests by enzyme-linked immunosorbent assay could be a feasible alternative for NPC screening in endemic areas. Further follow-up is needed to examine whether screening has an effect on decreasing mortality from NPC in these areas.
Assuntos
Anticorpos Antivirais , Antígenos Virais , Proteínas do Capsídeo , Detecção Precoce de Câncer/métodos , Neoplasias Nasofaríngeas/diagnóstico , Adulto , Carcinoma , China/epidemiologia , Ensaio de Imunoadsorção Enzimática , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Carcinoma NasofaríngeoRESUMO
While improving the performance on the out-of-distribution (OOD) benchmark dataset, the existing approach misses a portion of the valid discriminative information such that it reduces the performance on the same manifold OOD (SMOOD) data. The key to addressing this problem is to prompt the model to learn effective and comprehensive in-distribution (ID) semantic features. In this paper, two strategies are proposed to improve the generalization ability of the model to OOD data. Firstly, the original samples are replaced by features extracted from multiple "semantic perspectives" to obtain a comprehensive semantics of the samples; Second, the mean and variance of the batch samples are perturbed in the inference stage to improve the sensitivity of the model to the OOD data. The method we propose does not employ OOD samples, uses no pre-trained models in training, and does not require pre-processing of samples during inference. Experimental results show that our method enhances the semantic representation of ID data, surpasses state-of-the-art detection results on the OOD benchmark dataset, and significantly improves the performance of the model in detecting the SMOOD data.
RESUMO
In a prospective study, 42 048 adults residing in Zhongshan City, Guangdong, China, were followed for 16 years, and 171 of them developed nasopharyngeal carcinoma (NPC). Although Epstein-Barr virus (EBV) antibody levels of the cohort fluctuated, the antibody levels of 93% of the patients with NPC were raised and maintained at high levels for up to 10 years prior to diagnosis. This suggests that the serologic window affords an opportunity to monitor tumor progression during the preclinical stage of NPC development, facilitating early NPC detection. We reviewed the clinical records of the 171 patients with NPC in the prospective study to assess the efficacy of early NPC detection by serologic screening and clinical examination. Of the 171 patients, 51 had Stage I tumor (44 were among the 73 patients detected by clinical examination and 7 were among the 98 patients presented to outpatient department). Initial serologic screening predicted 58 (95.1%) of the 61 patients detected within 2 years. The risk of the screened population (58/3093) raised 13 times relative to cohort (61/42 048) during this period. Clinical examination detected all the 58 predicted cases, and 35 (60.3%) of which were diagnosed with Stage I tumor. The serologic prediction rate fell to 33.6% (37/110) 2 to 16 years after screening. The proportion of cases detected by clinical examination fell to 40.5% (15/37). The proportion of Stage I tumors among the cases detected by clinical examination during both periods remained at about 60%. We concluded that early detection of NPC can be accomplished by repeated serologic screening to maintain high prediction rates and by promptly examining screened subjects to detect tumors before the symptoms develop.
Assuntos
Anticorpos Antivirais/sangue , Antígenos Virais/imunologia , Proteínas do Capsídeo/imunologia , Carcinoma de Células Escamosas/diagnóstico , Detecção Precoce de Câncer/métodos , Neoplasias Nasofaríngeas/diagnóstico , Adulto , Idoso , Carcinoma de Células Escamosas/sangue , Carcinoma de Células Escamosas/patologia , Quimioterapia Adjuvante , Estudos de Coortes , Feminino , Herpesvirus Humano 4/imunologia , Humanos , Masculino , Pessoa de Meia-Idade , Neoplasias Nasofaríngeas/sangue , Neoplasias Nasofaríngeas/patologia , Metástase Neoplásica , Recidiva Local de Neoplasia , Estadiamento de Neoplasias , Indução de Remissão , Taxa de SobrevidaRESUMO
The tracking-by-detection framework receives growing attention through the integration with the convolutional neural networks (CNNs). Existing tracking-by-detection-based methods, however, fail to track objects with severe appearance variations. This is because the traditional convolutional operation is performed on fixed grids, and thus may not be able to find the correct response while the object is changing pose or under varying environmental conditions. In this paper, we propose a deformable convolution layer to enrich the target appearance representations in the tracking-by-detection framework. We aim to capture the target appearance variations via deformable convolution, which adaptively enhances its original features. In addition, we also propose a gated fusion scheme to control how the variations captured by the deformable convolution affect the original appearance. The enriched feature representation through deformable convolution facilitates the discrimination of the CNN classifier on the target object and background. The extensive experiments on the standard benchmarks show that the proposed tracker performs favorably against the state-of-the-art methods.
RESUMO
This paper proposes a computationally efficient method for traffic sign recognition (TSR). This proposed method consists of two modules: 1) extraction of histogram of oriented gradient variant (HOGv) feature and 2) a single classifier trained by extreme learning machine (ELM) algorithm. The presented HOGv feature keeps a good balance between redundancy and local details such that it can represent distinctive shapes better. The classifier is a single-hidden-layer feedforward network. Based on ELM algorithm, the connection between input and hidden layers realizes the random feature mapping while only the weights between hidden and output layers are trained. As a result, layer-by-layer tuning is not required. Meanwhile, the norm of output weights is included in the cost function. Therefore, the ELM-based classifier can achieve an optimal and generalized solution for multiclass TSR. Furthermore, it can balance the recognition accuracy and computational cost. Three datasets, including the German TSR benchmark dataset, the Belgium traffic sign classification dataset and the revised mapping and assessing the state of traffic infrastructure (revised MASTIF) dataset, are used to evaluate this proposed method. Experimental results have shown that this proposed method obtains not only high recognition accuracy but also extremely high computational efficiency in both training and recognition processes in these three datasets.
RESUMO
OBJECTIVE: To investigate the effect of paclitaxel and 5-flurouracil (5-Fu) on growth inhibition and apoptosis of human hepatoma BEL-7402 cells. METHODS: Growth inhibition of BEL-7402 cells treated with paclitaxel and 5-Fu, respectively, was measured by ATP-tumor chemosensitivity assay (ATP-TCA), and the cell cycle kinetics and apoptosis were analyzed by flow cytometry and microscopic examination. RESULTS: BEL-7402 cells were highly sensitive to paclitaxel with growth inhibition observed in both dose- and time-dependent manners (IC(50)=5.58 x 10(-7) mol/L). Paclitaxel induced significantly higher rate of cell apoptosis than the control group (P<0.05) but significantly lower rate than that induced by 5-Fu (P<0.01). Necrosis was observed predominantly in paclitaxel-treated cells whereas 5-Fu caused mainly cell apoptosis (P<0.05). Levels of apoptosis increased in proportion to the decrement of paclitaxel concentration but directly proportional to increment of 5-Fu concentration. CONCLUSIONS: Paclitaxel and 5-Fu are effective in inducing growth inhibition and apoptosis of BEL-7402 cells. While 5-Fu causes mainly apoptosis in hepatoma cells, the anticancer mechanism of paclitaxel is predominantly through induction of necrosis.
Assuntos
Apoptose/efeitos dos fármacos , Carcinoma Hepatocelular/patologia , Fluoruracila/farmacologia , Neoplasias Hepáticas/patologia , Paclitaxel/farmacologia , Antimetabólitos Antineoplásicos/farmacologia , Antineoplásicos Fitogênicos/farmacologia , Ciclo Celular/efeitos dos fármacos , Divisão Celular/efeitos dos fármacos , Humanos , Células Tumorais CultivadasRESUMO
OBJECTIVE: Nasopharyngeal Carcinoma (NPC) can be successfully treated by radiotherapy, if the tumor is confined to nasopharynx, but clinical onset is usually delayed to more advanced stages, when prognosis is poor. The objective is to determine efficacy of a new program for early NPC detection, which entails stratification of the NPC risk of target population according to serum levels of 3 Epstein Barr Virus (EBV) antibodies. METHOD: The sera of 1373 healthy adult residents from Zhongshan were collected and analyzed in this study from Mar 16, 2007 to Dec 31, 2007. The levels of EBNA1/IgA, zta/IgG and EBNA1/IgG were tested by ELISA. To stratify the subjects of 1373 adults into high, moderate and normal NPC risk groups by regression analysis of the levels of the EBV antibody. The high-risk groups of nasopharyngeal carcinoma risk could be followed-up every 3-6 month. RESULT: NPC risk of 1379 adults was stratified according to serum levels of the 3 EBV antibodies. Eleven (0.8%) were identified to be of high risk for NPC, having high levels of all three antibodies and/or IgA EBNA level > 3 rod. Clinical examination of high risk subjects detected 5 NPC cases, 3 cases detected in the first instance and 2 in follow-up examination 3 to 6 months hence. Three cases were diagnosed with UICC Stage I tumor (60%), one in the first instance and 2 in follow-up, and the 5 cases account for all NPC cases detected from the entire cohort over 28 months(100%). CONCLUSION: The new program affords an efficient and efficacious means for early NPC detection.
Assuntos
Detecção Precoce de Câncer/métodos , Neoplasias Nasofaríngeas/diagnóstico , Neoplasias Nasofaríngeas/epidemiologia , Anticorpos Antivirais/sangue , China/epidemiologia , Antígenos Nucleares do Vírus Epstein-Barr/imunologia , Humanos , Masculino , Pessoa de Meia-Idade , Triagem Multifásica , Neoplasias Nasofaríngeas/sangue , Neoplasias Nasofaríngeas/virologia , Medição de RiscoRESUMO
By extending integrated competition hypothesis, this paper presents an object-based visual attention model, which selects one object of interest using low-dimensional features, resulting that visual perception starts from a fast attentional selection procedure. The proposed attention model involves seven modules: learning of object representations stored in a long-term memory (LTM), preattentive processing, top-down biasing, bottom-up competition, mediation between top-down and bottom-up ways, generation of saliency maps, and perceptual completion processing. It works in two phases: learning phase and attending phase. In the learning phase, the corresponding object representation is trained statistically when one object is attended. A dual-coding object representation consisting of local and global codings is proposed. Intensity, color, and orientation features are used to build the local coding, and a contour feature is employed to constitute the global coding. In the attending phase, the model preattentively segments the visual field into discrete proto-objects using Gestalt rules at first. If a task-specific object is given, the model recalls the corresponding representation from LTM and deduces the task-relevant feature(s) to evaluate top-down biases. The mediation between automatic bottom-up competition and conscious top-down biasing is then performed to yield a location-based saliency map. By combination of location-based saliency within each proto-object, the proto-object-based saliency is evaluated. The most salient proto-object is selected for attention, and it is finally put into the perceptual completion processing module to yield a complete object region. This model has been applied into distinct tasks of robots: detection of task-specific stationary and moving objects. Experimental results under different conditions are shown to validate this model.
Assuntos
Algoritmos , Inteligência Artificial , Atenção , Biomimética/métodos , Modelos Teóricos , Reconhecimento Automatizado de Padrão/métodos , Robótica/métodos , Percepção Visual , Simulação por ComputadorAssuntos
Antineoplásicos/farmacologia , Carcinoma Hepatocelular/patologia , Neoplasias Hepáticas/patologia , Adenosina Trifosfatases/química , Ensaios de Seleção de Medicamentos Antitumorais , Feminino , Humanos , Luminescência , Medições Luminescentes , Masculino , Pessoa de Meia-Idade , Células Tumorais CultivadasRESUMO
OBJECTIVE: To study the effect of autogeneic PRP on proliferation and osteogenetic differentiation of human adipose-derived stem cells (ADSCs) in vitro. METHODS: ADSCs were isolated from adipose tissue obtained from donor undergoing liposuction and were cultured, and growth condition of the cells was observed by inverted microscope. ADSCs at passage 3 were cultured in adipogenic or chondrogenic medium and underwent identification, immunofluorescence staining observations for CD29 and CD44 were performed. ADSCs at passage 3 were divided into 2 groups: PRP group cultured by osteogenic induction culture medium containing 10 mL/L PRP, and control group cultured by osteogenic induction culture medium without PRP. Then growth condition of the cells was observed by inverted microscope. MTT method was used to observe cell proliferation activity 1, 2, 3, 4 and 5 days after culture. ALP activity detection was conducted 7, 14, 21 and 28 days after culture. ALP staining was performed on PRP group 7 and 14 days after culture. Alizarin red staining was performed on PRP group 14 days after culture to detect the formation of calcium nodule. RESULTS: Under the inverted microscope, most ADSCs at passage 3 were spindle-shaped and the doubling time was about 35 hours. Adipogenic and chondrogenic differentiation were confirmed, and the cells were positive for CD29 and CD44 immunofluorescence staining. MTT method revealed the absorbance value of PRP group at 1, 2, 3, 4 and 5 days was 0.137 +/- 0.015, 0.219 +/- 0.023, 0.367 +/- 0.031, 0.586 +/- 0.039 and 0.948 +/- 0.046, respectively, and in the control group, it was 0.081 +/- 0.009, 0.115 +/- 0.012, 0.162 +/- 0.017, 0.242 +/- 0.025 and 0.356 +/- 0.032, respectively, suggesting there were significant differences between two groups (P < 0.01). At 7 days after osteogenic induction, PRP group was positive for ALP staining, grey-black cell plasma and black precipitate were evident; the positive cells increased 14 days after osteogenic induction. ALP activity detection demonstrated the cell activity value of PRP group at 7, 14, 21 and 28 days was 23.96 +/- 2.05, 41.26 +/- 3.38, 38.12 +/- 3.03 and 35.89 +/- 2.24, respectively, while in the control group, it was 17.83 +/- 1.62, 26.64 +/- 2.37, 23.85 +/- 1.99 and 20.78 +/- 1.81, respectively, indicating there were significant differences between two groups (P < 0.01). At 14 days after osteogenic induction, Alizarin red staining showed the formation of calcium nodule in PRP group. CONCLUSION: Autogeneic PRP can improve the proliferation of human ADSCs and induce their osteogenic differentiation in vitro, providing a new source of seed cells for bone tissue engineering.