The redundant information, noise data generated in the process of single-modal feature extraction, and traditional learning algorithms are difficult to obtain ideal recognition performance. A multi-modal fusion emotion method for speech expressions based on deep is proposed. Firstly, corresponding extraction methods set up different single modalities. Among them, voice uses convolutional neural network-long short term memory (CNN-LSTM) network, facial expression video Inception-Res Net-v2...

This paper proposes an audio depression recognition method based on convolution neural network and generative antagonism model. First of all, preprocess the data set, remove long-term mute segments in splice rest into a new file. Then, features speech signal, such as Mel-scale Frequency Cepstral Coefficients (MFCCs), short-term energy spectral entropy, are extracted difference normalization algorithm. The matrix vector feature data, which represents unique attributes subjects' own voice, is...

Abstract Depth of invasion (DOI) can be calculated preoperatively by MRI, and whether MRI-determined DOI predict prognosis as well it used an indicator neck dissection in cT1N0 tongue squamous cell carcinoma (SCC) remains unknown. The main goal the current study was to answer these unknowns. A total 151 patients with surgically treated SCC were retrospectively enrolled, measured based on T1-weighted layers a 3.0T scan. Chi-square test evaluate association between clinical pathologic...

The purpose of this study is to explore the price transfer mechanism analysis and prediction domestic green certificate market based on convolutional neural network. By using network, time series economics methods, an innovative transmission model proposed, aiming deeply understand forecast trend. network method used analyze historical data certific ate in order reveal inherent law mechanism. Combined with analysis, paper analyzes trend periodicity change, provides reference for future Using...

The primary objective of this study is to evaluate the diagnostic efficacy gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA) -enhanced magnetic resonance imaging (MRI) in distinguishing breast cancer liver metastasis (BCLM) across different molecular types. Between August 2014 and July 2021, a cohort 270 patients histologically diagnosed with BCLM underwent examination through dynamic contrast-enhanced MRI (DCE-MRI). data collection encompassed clinical information...

The recovery of 3D models from multiple reference images involves not only the extraction shape, but also texture. Assuming that all surfaces are Lambertian, resulting final texture is typically computed as a linear combination textures. This is, however, optimal means for reconstructing textures, since this does model anisotropy in projection. Furthermore, spatial image sampling may be quite variable within fore-shortened surface. has important implications computer vision techniques...

Many applications based on Internet of Things (IoT) technology have recently founded in industry monitoring area. Thousands sensors with different types work together an system. Sensors at locations can generate streaming data, which be analyzed the data center. In this paper, we propose a framework for online sensor fault detection. We motivate our technique context problem value detection and event use Statistics Sliding Windows (SSW) to contain recent regress each window by Gaussian...

The multimedia big data in sensing and other IoT (Internet of Things) systems are high-volume, real-time, dynamic heterogeneous. These characteristics lead to new challenges security. When computation power resources some nodes very scarce, these become more serious that complex security process on is restricted by the aforementioned limited resources. Hence, confidentiality under constraints investigated this paper. Firstly, growth trend volume compared with computational discussed, an...

Abstract Rationale: Ameloblastoma is generally characterized as a benign tumor originating in odontogenic epithelium. However, few cases of metastatic malignant ameloblastoma have also been reported. Due to the low incidence ameloblastoma, there no established treatment regimen. To explore effective for we reported this case study. Patients concerns: This report described 28-year-old female patient with multiple metastasis (brain and lung). Diagnoses: The presented left mandible 2012. Three...

Fusobacterium nucleatum (F. nucleatum) often colonizes cancerous gastric tissues and is characterized by the promotion of platelet aggregation development visceral thrombosis. Venous thromboembolism (VTE) leads to a significant increase in mortality cancer (GC) patients. However, relationship between colonization F. prognosis GC patients still unknown.The aim this study was explore whether related complicated with VTE other potential risk factors.From 2017-2021, data 304 new VTEs during...

There exist many computer graphics techniques which could achieve high quality tree generation. However, only few works focus on realistic modeling of bark. Difficulties lie in the complex appearance bark surfaces from a single image. We address three main issues here: feature specification; height field assignment; and texture correction. For specification, we use texton channel analysis to specify variant common features, inluding ironbark, vertical horizontal fractures, tessellation,...

In this article, we study rates of convergence the generalization error multi-class margin classifiers. particular, develop an upper bound theory quantifying various large The permits a treatment general losses, convex or nonconvex, in presence absence dominating class. Three main results are established. First, for any fixed loss, there may be trade-off between ideal and actual performances with respect to choice class candidate decision functions, which is governed by approximation...

Transmission security of media data in video conference system has been a challenging task, since those are high volume and latency sensitive. Partial encryption other schemes effective approaches to encrypt data. However, partial disclosure can leak the secret about conference, moreover when wireless terminals joint systems, limited bandwidth computational capability will lead new challenges. To solve problems, novel scheme for with is proposed, which integrates three key modules...

Mutual trust has been the most important factor for users to do business with each other. But trustworthiness of an entire system provides assurance choose a particular e-commerce platform at first place. As difficult-to-observe property system, measurement overall is obstructed by quantification, accuracy and reliability. This paper proposes fuzzy hybrid multi-criteria analysis approach measure systems. Trust factors involved in systems are collected grouped. Then, these quantified placed...

Fusobacterium nucleatum (F. nucleatum) is a resident anaerobic bacterium, which in rare cases may invade blood from the head and neck or digestive tract to cause bacteremia induce venous thrombosis. F. closely related abdominal tumors, but it has not been reported relation renal tumors. We report herein possible case.This patient had kidney cancer with thrombosis right vein no sign of infection. After radical nephrectomy, thrombi formed his left vein, when removed, severe sepsis occurred. He...

The latest version of the National Comprehensive Cancer Network recommends neoadjuvant therapy followed by surgical treatment or radical chemoradiotherapy for patients with cT3N0M0. Neoadjuvant can improve prognosis locally advanced esophageal cancer. Therefore, evaluation prediction T stage is particularly important because could differently affect prognosis. Here, we establish a model to predict T2-3N0M0 help choose best strategy.From 1637 patents cancer, enrolled 48 and performed least...

Aim: To compare the apparent diffusion coefficient (ADC) derived from conventional diffusion-weighted imaging (DWI), and parameters intra-voxel incoherent motion (IVIM) diffusion, in differentiation of common focal malignant liver masses. Methods: DWI 53 patients with 85 masses (47 hepatocellular cancers [HCCs], 18 cholangiocarcinomas, 20 metastases) were recruited to receive a 3.0T MR scanner 2 b values (0, 700 s/mm [2] ) measure ADC 7 50, 100, 250, 500, 750, 1000 IVIM parameters: true (D),...

In the field of sequencing secret number, an important problem is how to establish efficient and secure protocol for sorting number. As a powerful tool in solving privacy problems, multipart computation more popular anonymous voting online auction. present study, related not many by far. order improve efficiency safety, we propose security based on homomorphic encryption. We also give analysis correctness highlight its feasibility.

Since media data in streaming system are high volume and latency sensitive, transmission security of them have been a challenge task. To solve the problem, partial encryption other schemes proposed. However those tradeoffs between efficiency. In multi-stream system, using one scheme to protect all streams will induce either unscalability or insecurity. this paper, idea is generalized from encrypting subset stream multi-streams. A mathematical model based dynamic optimal selective control...

The recording and playback system (RPS) in video conference needs to store mass of media data real-timely. Considering the security issue, should be encrypted before storing. Traditional full encryption partial algorithms are not applicable RPS because they could adjust their speed meet throughput variation real-time RPS. In this paper, a novel lightweight speed-adaptive media-data (SAME) scheme is proposed firstly. Secondly, SAME improved packet-based algorithm according implementation...