Improved Environment-Aware-Based Noise Reduction System for Cochlear Implant Users Based on a Knowledge Transfer Approach: Development and Usability Study.

BACKGROUND: Cochlear implant technology is a well-known approach to help deaf individuals hear speech again and can improve speech intelligibility in quiet conditions; however, it still has room for improvement in noisy conditions. More recently, it has been proven that deep learning-based noise reduction, such as noise classification and deep denoising autoencoder (NC+DDAE), can benefit the intelligibility performance of patients with cochlear implants compared to classical noise reduction algorithms. OBJECTIVE: Following the successful implementation of the NC+DDAE model in our previous study, this study aimed to propose an advanced noise reduction system using knowledge transfer technology, called NC+DDAE_T; examine the proposed NC+DDAE_T noise reduction system using objective evaluations and subjective listening tests; and investigate which layer substitution of the knowledge transfer technology in the NC+DDAE_T noise reduction system provides the best outcome. METHODS: The knowledge transfer technology was adopted to reduce the number of parameters of the NC+DDAE_T compared with the NC+DDAE. We investigated which layer should be substituted using short-time objective intelligibility and perceptual evaluation of speech quality scores as well as t-distributed stochastic neighbor embedding to visualize the features in each model layer. Moreover, we enrolled 10 cochlear implant users for listening tests to evaluate the benefits of the newly developed NC+DDAE_T. RESULTS: The experimental results showed that substituting the middle layer (ie, the second layer in this study) of the noise-independent DDAE (NI-DDAE) model achieved the best performance gain regarding short-time objective intelligibility and perceptual evaluation of speech quality scores. Therefore, the parameters of layer 3 in the NI-DDAE were chosen to be replaced, thereby establishing the NC+DDAE_T. Both objective and listening test results showed that the proposed NC+DDAE_T noise reduction system achieved similar performances compared with the previous NC+DDAE in several noisy test conditions. However, the proposed NC+DDAE_T only required a quarter of the number of parameters compared to the NC+DDAE. CONCLUSIONS: This study demonstrated that knowledge transfer technology can help reduce the number of parameters in an NC+DDAE while keeping similar performance rates. This suggests that the proposed NC+DDAE_T model may reduce the implementation costs of this noise reduction system and provide more benefits for cochlear implant users.

Bayesian Edge Detector Using Deformable Directivity-Aware Sampling Window.

Conventional image entropy merely involves the overall pixel intensity statistics which cannot respond to intensity patterns over spatial domain. However, spatial distribution of pixel intensity is definitely crucial to any biological or computer vision system, and that is why gestalt grouping rules involve using features of both aspects. Recently, the increasing integration of knowledge from gestalt research into visualization-related techniques has fundamentally altered both fields, offering not only new research questions, but also new ways of solving existing issues. This paper presents a Bayesian edge detector called GestEdge, which is effective in detecting gestalt edges, especially useful for forming object boundaries as perceived by human eyes. GestEdge is characterized by employing a directivity-aware sampling window or mask that iteratively deforms to probe or explore the existence of principal direction of sampling pixels; when convergence is reached, the window covers pixels best representing the directivity in compliance with the similarity and proximity laws in gestalt theory. During the iterative process based on the unsupervised Expectation-Minimization (EM) algorithm, the shape of the sampling window is optimally adjusted. Such a deformable window allows us to exploit the similarity and proximity among the sampled pixels. Comparisons between GestEdge and other edge detectors are shown to justify the effectiveness of GestEdge in extracting the gestalt edges.

Fast Visual Tracking Based on Convolutional Networks.

Recently, an upsurge of deep learning has provided a new direction for the field of computer vision and visual tracking. However, expensive offline training time and the large number of images required by deep learning have greatly hindered progress. This paper aims to further improve the computational performance of CNT which is reported to deliver 5 fps performance in visual tracking, we propose a method called Fast-CNT which differs from CNT in three aspects: firstly, an adaptive k value (rather than a constant 100) is determined for an input video; secondly, background filters used in CNT are omitted in this work to save computation time without affecting performance; thirdly, SURF feature points are used in conjunction with the particle filter to address the drift problem in CNT. Extensive experimental results on land and undersea video sequences show that Fast-CNT outperforms CNT by 2~10 times in terms of computational efficiency.

[Preparation and in vitro evaluation of ampelopsin-loaded nanomicelles].

To improve the solubility and antitumor activity of ampelopsin, ampelopsin-loaded nanomicelles from the mixture of pluronic F127 and D-α-tocopheryl polyethylene glycol 1000 succinate (TPGS1000) were prepared by film-thin hydration method, in order to optimize the process conditions and physicochemical properties. The antitumor activities against MCF-7 cells between ampelopsin and nanomicelles were compared by MTT method, respectively. The results showed that the optimal nanomicelles were round with the nanometric size of (22.6±0.5) nm, encapsulation efficiency rate of (80.42±1.13)%, and drug-loading rate of (4.41±0.26)%. The solubility of ampelopsin in mixed nanomicelles significantly increased by 16 times. In different release media, the mixed nanomicelles could release more than 90% of drug in 8 h, and showed stronger cytotoxicity and inhibition against MCF-7 cells (P<0.01). The mixed nanomicelles can be used as new drug delivery system of ampelopsin.

Epidemiologic characterization of 30 confirmed cases of human infection with avian influenza A(H7N9) virus in Hangzhou, China.

BACKGROUND: We examined the clinical and epidemiological characteristics of 30 cases of human infection with avian influenza A(H7N9) virus in Hangzhou and investigated their external environments to provide evidence for contact tracing and disease prevention and control. METHODS: The cases confirmed from April 1 through May 1, 2013 were studied. Field epidemiologic surveys were conducted to collect the clinical and epidemiologic data. Case-related and environmental specimens were collected for etiologic detection. RESULTS: Thirty cases of human infection with avian influenza A(H7N9) virus were confirmed in Hangzhou from April 1 through May 1, 2013, including one pregnant woman and three deaths. The median age of the patients was 62 years (range: 38-86 years). Twenty-three of the patients were men (76.67%). The median duration between disease onset and occurrence of respiratory failure and confirmed diagnosis was 5 and 6 days, respectively. Maximum medical observation of 666 close contacts of the patients revealed no irregularity. Of 314 external environmental specimens, the overall positive detection rate of H7N9 nucleic acid was 28.98%. Eight districts of Hangzhou city had positive detections in the external environments, the highest rate being in Yuhang District (78.13%). Statistical analysis of the specimen collection locations indicates a significant difference between the case-linked locations and the non-case locations (χ2 = 16.563, p < 0.05) in terms of H7N9 viral nucleic acid detection rate. No epidemiologic link has been found among the 30 cases. CONCLUSIONS: Most of the infected were retired individuals aged 60 years or older. Men made the majority. The cases are sporadic at present, with no evidence of human-to-human transmission. Exposures to poultry and live poultry markets may be important sources of infection.

Pandemic influenza A(H1N1) 2009 virus in pregnancy.

Two hundred fourteen abstracts and 87 full texts regarding pregnant women infected with pandemic influenza A(H1N1) 2009 virus were systematically reviewed by using a PubMed search and assessing pandemic, clinical, laboratory test, vaccine, and control experiences. Both policy and health education were excluded. This review counted the total number of pregnant cases from different countries and analyzed their epidemic features, including trimester distribution, morbidity, hospitalization, intensive care unit admissions, maternal mortality, underlying diseases, complications, high-risk factors for death, pregnancy outcome, and clinical symptoms compared with the previous pandemic seasonal influenza A/H1N1 as compared with the general population. Early identification and treatment were the most important factors in different countries and areas examined. The vaccine and antiviral drugs that have been the most efficient means to control the novel virus appear to be safe but require more extensive study. In the future, the focus should be placed on understanding vertical transmission and the severe mechanisms.

Two adenovirus serotype 3 outbreaks associated with febrile respiratory disease and pharyngoconjunctival fever in children under 15 years of age in Hangzhou, China, during 2011.

Adenovirus serotype 3 and 7 outbreaks have occurred periodically in northern, eastern, and southern China since 1955, but there has been no report since the adenovirus serotype 7 outbreak first occurred in Hangzhou, China, in 1991. Here we explored the epidemiology and etiology of two adenovirus serotype 3 outbreaks in Hangzhou in 2011. One acute respiratory outbreak was found in Chun'an County, where a total of 371 cases were confirmed in 5 of 23 towns from 4 to 31 May 2011. The outbreak affected 18.57% (13/70) of schools and 14.49% (90/621) of classes. The incidence was 5.18% (371/7,163). The population was distributed among individuals ages 7 to 15 years. No parents or teachers were infected. Another pharyngoconjunctival fever outbreak was discovered in the Chenjinglun Swimming Center located in the Xihu District between 1 and 15 July 2011. A total of 134 cases were confirmed in 900 amateur swimmers, with an incidence of 14.89% (134/900). The ages ranged from 4 to 9 years. The two outbreaks had no severe complications or death. The viruses in 66.67% (10/15) of throat swabs from children with acute respiratory infections and 100% (10/10) of the swabs from children with pharyngoconjunctival fever were confirmed to be adenovirus serotype 3 with 100% homology by PCR. Of these samples, 60.0% (12/20) had a classical characteristic cytopathic effect, presented as grape-like clusters at 72 h after infection in HEp-2 cells. In conclusion, the acute respiratory infection and pharyngoconjunctival fever outbreak in Hangzhou were caused by the completely homologous type 3 adenovirus in subgenus B. Moreover, these outbreaks demonstrated rapid transmission rates, possibly due to close contact and droplet transmission.

Bio-inspired contour extraction via EM-driven deformable and rotatable directivity-probing mask.

This paper presents a novel bio-inspired edge-oriented approach to perceptual contour extraction. Our method does not rely on segmentation and can unsupervised learn to identify edge points that are readily grouped, without invoking any connecting mechanism, into object boundaries as perceived by human. This goal is achieved by using a dynamic mask to statistically assess the inter-edge relations and probe the principal direction that acts as an edge-grouping cue. The novelty of this work is that the mask, centered at a target pixel and driven by EM algorithm, can iteratively deform and rotate until it covers pixels that best fit the Bayesian likelihood of the binary class w.r.t a target pixel. By creating an effect of enlarging receptive field, contiguous edges of the same object can be identified while suppressing noise and textures, the resulting contour is in good agreement with gestalt laws of continuity, similarity and proximity. All theoretical derivations and parameters updates are conducted under the framework of EM-based Bayesian inference. Issues of stability and parameter uncertainty are addressed. Both qualitative and quantitative comparison with existing approaches proves the superiority of the proposed method in terms of tracking curved contours, noises/texture resilience, and detection of low-contrast contours.

2009 pandemic characteristics and controlling experiences of influenza H1N1 virus 1 year after the inception in Hangzhou, China.

This paper described the epidemiology and controlling experiences of influenza H1N1 in Hangzhou in the past 1 year. A total of 2,078 cases confirmed by real-time quantitative PCR till March 31, 2010, were analyzed by SPSS 12.0 software. During the early pandemic stage, a patient must be tested for H1N1 nucleic acid once he/she had influenza-like symptoms with the epidemiological history in 7 days, and be diagnosed if it was positive. But in the pandemic peak, we made efforts to identify and save severe cases combined with pneumonia or hypoxemia or respiratory failure or septic shock or multiple organ dysfunctions and failure. In general, the prevalence was 2.77/100,000 (2,078/7,510,844); severity rate, 10.44% (217/2,078); fatality rate, 0.48% (10/2,078). The carrier and secondary attack rate were 9.52% (58/612) and 8.66% (53/612), respectively. About 50% of serious cases and 100% of deaths had the basic underlying diseases: cardiovascular diseases, 13.66% (25/217); chronic lung disease, 12.02% (22/217); pregnant, 7.1% (13/217). Of all cases aged from 1 month to 89 years, 52.99% (1,435/2,708) were in the 10-29 years, with most of them distributed in downtown area. The timeline showed two epidemic peaks occurred in September and November 2009, respectively. Furthermore, the hemagglutinin gene remained 99% identical with the American and vaccine strains, but only 70% with the 1947-2008 Chinese strains. In conclusion, Hangzhou pandemic influenza H1N1 was caused by the highly conserved virus, with low prevalence, transmission, and mortality, because we took efficient controlling methods.

Retrospective study of risk factors for mortality in human avian influenza A(H7N9) cases in Zhejiang Province, China, March 2013 to June 2014.

BACKGROUND: The influenza A(H7N9) virus causes a serious disease that threatens human health. Fatalities associated with human infections caused by this virus are of great public health concern; however, the possible risk factors are not yet fully known. METHODS: A stratified sampling method, incorporating household income levels and a random number table method, was used to select laboratory-confirmed A(H7N9) cases for this study. Eighty-five patients were selected randomly from 139 laboratory-confirmed A(H7N9) cases occurring in Zhejiang Province between March 1, 2013 and June 30, 2014. Data were collected using a standard method. To test the statistical significance among discrete variables, univariate analyses were used to compare two groups. The Kaplan-Meier product-limit method was used to analyze the patient survival fraction. The Cox proportional hazards regression model was used to analyze all variables with p ≤ 0.05 in the univariate analysis. Lastly, a stepwise procedure was used to construct a final model with a significance level of p > 0.10 for removal and p<0.05 for re-entry. RESULTS: A total of 85 patients with H7N9 virus infection were identified. Among these, 30 (35.29%) died. In the univariate analysis, the following factors were associated with a high risk of influenza A(H7N9) case fatality: age ≥ 60 years (p=0.008), low education level (p=0.030), chronic diseases (p=0.029), poor hand hygiene (p=0.010), time from illness onset to the first medical visit (p=0.029) and to intensive care unit admission (p=0.008), an incubation period of ≤ 5 days (p=0.039), a peak C-reactive protein ≥ 120 mg/l (p=0.012), increased initial neutrophil count (p=0.020), decreased initial lymphocyte count (p=0.021), and initial infection of both lungs (p=0.003). Multivariate analysis confirmed that the independent predictors of H7N9 virus infection mortality in Zhejiang, China were hand hygiene (hazard ratio (HR) 5.163, 95% confidence interval (CI) 1.164-22.661), age (HR 1.042, 95% CI 1.007-1.076), and peak CRP (HR 1.009, 95% CI 1.002-1.016). CONCLUSIONS: Improvements in immunity, early case identification and treatment, and personal protection measures are key to addressing the high human avian influenza A(H7N9) case fatality rate.

All-trans retinoic acid stealth liposomes prevent the relapse of breast cancer arising from the cancer stem cells.

The relapse of cancer is mostly due to the proliferation of cancer stem cells which could not be eliminated by a standard chemotherapy. A new kind of all-trans retinoic acid stealth liposomes was developed for preventing the relapse of breast cancer and for treating the cancer in combination with a cytotoxic agent, vinorelbine stealth liposomes. In vitro studies were performed on the human breast cancer MCF-7 and MDA-MB-231 cells. In vivo evaluations were performed on the newly established relapse model with breast cancer stem cells. Results showed that the particle size of all-trans retinoic acid stealth liposomes was approximately 80nm, and the encapsulation efficiency was >90%. Breast cancer stem cells were identified with the CD44(+)/CD24(-) phenotype and characterized with properties: resistant to cytotoxic agent, stronger capability of proliferation, and stronger capability of differentiation. Inhibitory effect of all-trans retinoic acid stealth liposomes was more potent in cancer stem cells than in cancer cells. The mechanisms were defined to be two aspects: arresting breast cancer stem cells at the G(0)/G(1) phase in mitosis, and inducing the differentiation of breast cancer stem cells. The cancer relapse model was successfully established by xenografting breast cancer stem cells into NOD/SCID mice, and the formation and growth of the xenografted tumors were significantly inhibited by all-trans retinoic acid stealth liposomes. The combination therapy of all-trans retinoic acid stealth liposomes with vinorelbine stealth liposomes produced the strongest inhibitory effect to the relapse tumor model. It could be concluded that all-trans retinoic acid stealth liposomes could be used for preventing the relapse of breast cancer by differentiating cancer stem cells and arresting the cell-cycle, and for treating breast cancer as a co-therapy, thus providing a novel strategy for treating breast cancer and preventing relapse derived from breast cancer stem cells.