Microphone array based methods are gradually applied in the front-end speech enhancement and speech recognition improvement for cochlear implant in recent years. By placing several microphones in different locations in space, this method can collect multi-channel signals containing a lot of spatial position and orientation information. Microphone array can also yield specific beamforming mode to enhance desired signal and suppress ambient noise, which is particularly suitable to be applied in face-to-face conversation for cochlear implant users. And its application value has attracted more and more attention from researchers. In this paper, we describe the principle of microphone array method, analyze the microphone array based speech enhancement technologies in present literature, and further present the technical difficulties and development trend.
In order to fully explore the neural oscillatory coupling characteristics of patients with mild cognitive impairment (MCI), this paper analyzed and compared the strength of the coupling characteristics for 28 MCI patients and 21 normal subjects under six different-frequency combinations. The results showed that the difference in the global phase synchronization index of cross-frequency coupling under δ-θ rhythm combination was statistically significant in the MCI group compared with the normal control group (P = 0.025, d = 0.398). To further validate this coupling feature, this paper proposed an optimized convolutional neural network model that incorporated a time-frequency data enhancement module and batch normalization layers to prevent overfitting while enhancing the robustness of the model. Based on this optimized model, with the phase locking value matrix of δ-θ rhythm combination as the single input feature, the diagnostic accuracy of MCI patients was (95.49 ± 4.15)%, sensitivity and specificity were (93.71 ± 7.21)% and (97.50 ± 5.34)%, respectively. The results showed that the characteristics of the phase locking value matrix under the combination of δ-θ rhythms can adequately reflect the cognitive status of MCI patients, which is helpful to assist the diagnosis of MCI.
ObjectiveTo compare the effectiveness of T2 weighted image (T2WI) and some compounded MRI techniques, including T2WI combined with magnetic resonance spectroscopy (T2WI+MRS), T2WI combined with diffusion weighted imaging (T2WI+DWI) and T2WI combined with dynamic contrast-enhancement [T2WI+(DCE-MRI)] respectively, with 1.5 T MR scanner in diagnosing prostate cancer through a blinding method. MethodsBetween March 2011 and April 2013, two observers diagnosed 59 cases with a blinding method. The research direction of radiologist A was to diagnose prostate cancer. The observers diagnosed and scored the cases with T2WI, T2WI+(DCE-MRI), T2WI+MRS, T2WI+DWI and compositive method respectively. The data were statistically analyzed with receiver operating characteristic (ROC) curve. ResultsAccording to the ROC curve, both observers got the sequence of area under curve (AUC) as T2WI+DWI > T2WI+(DCE-MRI) > T2WI+MRS > T2WI. On the basis of the result from observer A, the AUC from each technique was similar. The AUC of T2+DWI was slightly bigger than others. The specificity of single T2WI was the lowest; the sensitivity of T2WI was slightly higher. The AUC of the compositive method was marginally larger than T2WI+DWI. According to the result from observer B, the AUC of T2WI+DWI was obviously larger than the others. The AUC of single T2WI was much smaller than the other techniques. The single T2WI method had the lowest sensitivity and the highest specificity. The AUC of T2WI+DWI was slightly larger than the compositive method. The AUC of T2WI+(DCE-MRI), T2WI+MRS, single T2WI methods from observer A was obviously higher than those from the score of observer B. The AUC of T2WI+DWI from the two observers was similar. ConclusionThe method of combined T2WI and functional imaging sequences can improve the diagnosing specificity when a 1.5 T MR scanner is used. T2WI+DWI is the best method in diagnosing prostate cancer with least influence from the experience of observers in this research. The compositive method can improve the diagnosis of prostate cancer effectively, but when there are contradictions between different methods, the T2WI+DWI should be considered as a key factor.
Parkinson’s disease patients have early vocal cord damage, and their voiceprint characteristics differ significantly from those of healthy individuals, which can be used to identify Parkinson's disease. However, the samples of the voiceprint dataset of Parkinson's disease patients are insufficient, so this paper proposes a double self-attention deep convolutional generative adversarial network model for sample enhancement to generate high-resolution spectrograms, based on which deep learning is used to recognize Parkinson’s disease. This model improves the texture clarity of samples by increasing network depth and combining gradient penalty and spectral normalization techniques, and a family of pure convolutional neural networks (ConvNeXt) classification network based on Transfer learning is constructed to extract voiceprint features and classify them, which improves the accuracy of Parkinson’s disease recognition. The validation experiments of the effectiveness of this paper’s algorithm are carried out on the Parkinson’s disease speech dataset. Compared with the pre-sample enhancement, the clarity of the samples generated by the proposed model in this paper as well as the Fréchet inception distance (FID) are improved, and the network model in this paper is able to achieve an accuracy of 98.8%. The results of this paper show that the Parkinson’s disease recognition algorithm based on double self-attention deep convolutional generative adversarial network sample enhancement can accurately distinguish between healthy individuals and Parkinson’s disease patients, which helps to solve the problem of insufficient samples for early recognition of voiceprint data in Parkinson’s disease. In summary, the method effectively improves the classification accuracy of small-sample Parkinson's disease speech dataset and provides an effective solution idea for early Parkinson's disease speech diagnosis.
Aiming at the problems of missing important features, inconspicuous details and unclear textures in the fusion of multimodal medical images, this paper proposes a method of computed tomography (CT) image and magnetic resonance imaging (MRI) image fusion using generative adversarial network (GAN) and convolutional neural network (CNN) under image enhancement. The generator aimed at high-frequency feature images and used double discriminators to target the fusion images after inverse transform; Then high-frequency feature images were fused by trained GAN model, and low-frequency feature images were fused by CNN pre-training model based on transfer learning. Experimental results showed that, compared with the current advanced fusion algorithm, the proposed method had more abundant texture details and clearer contour edge information in subjective representation. In the evaluation of objective indicators, QAB/F, information entropy (IE), spatial frequency (SF), structural similarity (SSIM), mutual information (MI) and visual information fidelity for fusion (VIFF) were 2.0%, 6.3%, 7.0%, 5.5%, 9.0% and 3.3% higher than the best test results, respectively. The fused image can be effectively applied to medical diagnosis to further improve the diagnostic efficiency.
ObjectiveTo investigate the diagnostic performance of parameters of arterial enhancement fraction (AEF) based on enhanced CT with histogram analysis in the severity of liver cirrhosis.MethodsThe patients with liver cirrhosis clinically confirmed and met the inclusion criteria were included from January 2016 to December 2018 in the First Affiliated Hospital of Chengdu Medical College, then them were divided into grade A, B, and C according to the Child-Pugh score. Meanwhile, the patients without liver disease were selected as the control group. All patients underwent the upper abdomen enhanced CT scan with three-phase and the biochemical examination of liver function. The parameters of AEF histogram were obtained by using the CT Kinetics software, and the aspartic aminotransferase and platelet ratio index (APRI) was calculated. The differences of parameters of AEF histogram and APRI among these patients with liver cirrhosis and without liver disease were analyzed. The diagnostic performance was evaluated by using the area under curve (AUC) of receivers operating characteristic curve.ResultsEighty-five patients with liver cirrhosis were included in this study, including 25, 41, and 19 patients with grade A, B, and C of Child-Pugh score, respectively, and there were 20 patients in the control group. The consistencies in measuring the parameters of AEF histogram twice for the same observer and between the two observers were good (intraclass correlation coefficient was 0.938 and 0.907, respectively). The mean, median, and kurtosis of AEF histogram and the APRI among the grade A, B, C of Child-Pugh score, and control group had significant differences (all P<0.001) and these indexes were positively correlated with the severity of liver cirrhosis (rs=0.811, P<0.001; rs=0.827, P<0.001; rs=0.731, P<0.001; rs=0.711, P<0.001). The AUC of the mean, median, kurtosis, and APRI in diagnosing grade A of liver cirrhosis was 0.829, 0.841, 0.747, and 0.718, respectively; which in diagnosing grade B of liver cirrhosis was 0.847, 0.734, 0.704, and 0.736, respectively; in diagnosing grade C of liver cirrhosis was 0.646, 0.825, 0.782, and 0.853, respectively.ConclusionThe mean and median of AEF histogram parameters based on enhanced CT with three-phase and serological APRI are useful in diagnosis of grage A, B, and C of liver cirrhosis, respectively.
High-frequency steady-state asymmetric visual evoked potential (SSaVEP) provides a new paradigm for designing comfortable and practical brain-computer interface (BCI) systems. However, due to the weak amplitude and strong noise of high-frequency signals, it is of great significance to study how to enhance their signal features. In this study, a 30 Hz high-frequency visual stimulus was used, and the peripheral visual field was equally divided into eight annular sectors. Eight kinds of annular sector pairs were selected based on the mapping relationship of visual space onto the primary visual cortex (V1), and three phases (in-phase[0º, 0º], anti-phase [0º, 180º], and anti-phase [180º, 0º]) were designed for each annular sector pair to explore response intensity and signal-to-noise ratio under phase modulation. A total of 8 healthy subjects were recruited in the experiment. The results showed that three annular sector pairs exhibited significant differences in SSaVEP features under phase modulation at 30 Hz high-frequency stimulation. And the spatial feature analysis showed that the two types of features of the annular sector pair in the lower visual field were significantly higher than those in the upper visual field. This study further used the filter bank and ensemble task-related component analysis to calculate the classification accuracy of annular sector pairs under three-phase modulations, and the average accuracy was up to 91.5%, which proved that the phase-modulated SSaVEP features could be used to encode high- frequency SSaVEP. In summary, the results of this study provide new ideas for enhancing the features of high-frequency SSaVEP signals and expanding the instruction set of the traditional steady state visual evoked potential paradigm.
Speech enhancement methods based on microphone array adopt many microphones to record speech signal simultaneously. As spatial information is increased, these methods can increase speech recognition for cochlear implant in noisy environment. Due to the size limitation, the number of microphones used in the cochlear implant cannot be too large, which limits the design of microphone array beamforming. To balance the size limitation of cochlear implant and the spatial orientation information of the signal acquisition, we propose a speech enhancement and beamforming algorithm based on dual thin uni-directional / omni-directional microphone pairs (TP) in this paper. Each TP microphone contains two sound tubes for signal acquisition, which increase the overall spatial orientation information. In this paper, we discuss the beamforming characteristics with different gain vectors and the influence of the inter-microphone distance on beamforming, which provides valuable theoretical analysis and engineering parameters for the application of dual microphone speech enhancement technology in cochlear implants.
During the automatic reconstruction of panoramic images, the effect of dental arch curve fitting will affect the integrity of the content of the panoramic image. Metal implants in the patient’s mouth usually lead to a decrease in the contrast of the panoramic image, which affects the doctor’s diagnosis. In this paper, an automatic oral panoramic image reconstruction method was proposed. By calculating key image areas and image extraction fusion algorithms, the dental arch curve could be automatically detected and adjusted on a small number of images, and the intensity distribution of teeth, bone tissue and metal implants on the image could be adjusted to reduce the impact of metal on other tissues, to generate high-quality panoramic images. The method was tested on 50 cases of cone beam computed tomography (CBCT) data with good results, which can effectively improve the quality of panoramic images.
Objective To summarize ultrasonography, CT and (or) MRI imaging features of cystic liver lesions so as to improve its diagnostic accuracy. Methods The literatures relevant imaging studies of different types of cystic liver lesions at home and abroad were searched. Then with the etiology as clue, the imaging fetures of ultrasonography, CT and (or) MRI plain scan and enhancement scan were summarized. Results The cystic liver lesions had many types, their imaging findings were different and existed overlaps. The diagnosis and differential diagnosis of atypical cases were difficult. ① For the simple hepatic cyst, it was a round cystic mass with water-like echo, density and signal. The boundary was clear, and there was no separation in the cyst, without contrast enhancement. The sensitivity and specificity of diagnosing were higher by ultrasonography and MRI as compared with CT. ② For the bile duct hamartoma and Caroli diease, they were manifested as multiple cysts, widely distributed in the whole liver, without enhancement for the most lesions. The multiple cystic lesions without communicating with the bile duct was the key sign of differential diagnosis for these two dieases. ③ Enhancing mural nodules were more common in cystadenocarcinoma than cystadenoma. The accurate diagnosis of biliary cystadenoma depended on combination of ultrasonography, CT, and MRI findings. ④ For the cystic liver metastatic tumor, it was multiple cystic neoplasms in the liver parenchyma or around the liver. CT was the main method for the diagnosis, and which showed that the density was lower than that of the liver parenchyma, peripheral ring-enhanced lesion as enhanced scan. It was easy to distinguish with simple hepatic cyst by MRI. ⑤ For the cystic hepatocellular carcinoma, it presented as a multilocular cystic solid tumor. The presence of tumor thrombus in portal vein could help to the diagnosis. ⑥ For the undifferentiated embryonal sarcoma, CT plain scan showed the cystic low density mass with clear boundary, the edge with calcification, enhanced scan showed that the soft tissue composition presented continuous strengthening sign. There was no specific signal in MRI plain scan, and the periphery of the tumor was slowly strengthening. ⑦ For the liver abscess, it was easy to diagnose because it had different characteristic features in different pathological phase, but it was misdiagnosis of intrahepatic cholangiocarcinoma when its symptoms were atypical. ⑧ The ultrasonography and the CT were the optimal methods for the hepatic cystic echinococcosis and the hepatic alveolar echinococcosis respectively. The significances of imaging were to determine the activity of hydatid cyst and to identify anatomy structure among alveolar echinococcosis, bile duct and blood vessel, and judge invasion or not, MRCP was important for diagnosis. Conclusions Abdominal ultrasonography could be used as the first choice for diagnosis of cystic liver lesions, CT and MRI could be used as effective supplementary methods for it. A combination of various imaging techniques is key to diagnosis. Moreover, number and morphology of lesion, and solid component or not are important imaging features of diagnosis and differential diagnosis of cystic liver lesion.