Object detection is a fundamental task in computer vision that involves accurately locating and classifying objects within images or video frames. In remote sensing, this particularly challenging due to the high resolution, multi-scale features, diverse ground object characteristics inherent satellite UAV imagery. These challenges necessitate more advanced approaches for effective such environments. While deep learning methods have achieved remarkable success sensing detection, they...

In a complex electromagnetic environment, there are cases where the noise is uncertain and difficult to estimate, which poses great challenge spectrum sensing systems. This paper proposes cooperative method based on empirical mode decomposition information geometry. The mainly includes two modules, signal feature extraction module K‐medoids. module, firstly, modal algorithm used denoise signals collected by secondary users, so as reduce influence of subsequent process. Further, problem...

Medical text learning has recently emerged as a promising area to improve healthcare due the wide adoption of electronic health record (EHR) systems. The complexity medical such diverse length, mixed types, and full jargon, poses great challenge for developing effective deep models. BERT presented state-of-the-art results in many NLP tasks, classification question answering. However, standalone model cannot deal with text, especially lengthy clinical notes. Herein, we develop new called...

Deep learning based wideband (16kHz) acoustic echo cancellation (AEC) approaches have surpassed traditional methods. This work proposes a deep hierarchical fusion (DHF) network with intra-network and inter-network to further improve the AEC performance. Meanwhile, this extends existing systems enable fullband (48kHz) while simultaneously ensuring automatic speech recognition compatibility by incorporating an ASR loss. The proposed system has ranked 2nd place in ICASSP 2022's Challenge.

This study investigates the method of analyzing emotional tendencies in music courses and its application lesson plan evaluation. Using a weighted to analyze curriculum, compares results with existing literature, demonstrating superior accuracy proposed method. To evaluate quality, combination self-assessment, mutual evaluation, group middle school evaluation form is recommended for comprehensive assessment. The study's comment polarity achieves an rate 69.19%, significantly outperforming...

Speech bandwidth extension (BWE) has demonstrated promising performance in enhancing the perceptual speech quality real communication systems. Most existing BWE researches primarily focus on fixed upsampling ratios, disregarding fact that effective of captured audio may fluctuate frequently due to various capturing devices and transmission conditions. In this paper, we propose a streaming adaptive solution dubbed BAE-Net, which is suitable handle low-resolution with unknown varying...

Purpose Turnover intention is a critical predictor of an employee’s turnover behaviour. A high level rate significantly affects the productivity and morale enterprise. Previous research has indicated that job satisfaction plays role in influencing employee's intention, but underlying factors related to remain under-explored, which impedes development effective strategies for reducing intention. In addition, little examined context COVID-19 pandemic, specifically Chinese construction...

This paper describes our submission to the fourth Acoustic Echo Cancellation (AEC) Challenge, which is part of ICASSP 2023 Signal Processing Grand Challenge. The proposed system developed based on earlier submitted 2022 AEC challenge with significant latency and network structure improvement, while achieving better subjective results.

Multinomial logit bandit is a sequential subset selection problem which arises in many applications. In each round, the player selects K-cardinality from N candidate items, and receives reward governed by multinomial (MNL) choice model considering both item utility substitution property among items. The player's objective to dynamically learn parameters of MNL maximize cumulative over finite horizon T. This faces exploration-exploitation dilemma, involved combinatorial nature makes it...

The paper presents a cooperative MAC based on opportunistic relaying for ad hoc networks. In the proposed MAC, scheme will be activated condition that source contended channel successful, but destination receives inaccurate packets because of noise. Based relaying, neighbor which has best to selected as relay help transmitting packets. An innovative analytical model considers both and is presented evaluate performance MAC. Simulations are also performed validate model. comparison with DCF...

In Semantic Web, modeling knowledge graph based on RDF becomes more and popular. There is quite a lot of spatiotemporal information in recent works focus not only general data but also data. Existing efforts are mainly to add labels RDF, which expand triple into quad or quintuple. However, extra often cause additional overhead for the system lead inefficient organization management. order overcome this limitation, we propose an stRDFS model by labeling properties with features corresponding...

This paper propose improved support vector machine algorithm. The algorithm includes preprocessing the sample training set, improvement of binary tree classification and incremental learning Considering specific precision requirements analog circuit fault diagnosis, three algorithms are integrated, achieve good results. simulation demonstrate that has higher faster diagnosis speed compared to traditional

Modified-DFT (MDFT) filter banks permit subchannels with linear phase characteristics, and provide high degrees of computational efficiency. However, in MDFT exhibiting narrow transition-bandwidths, the length prototype becomes prohibitively long, reducing It is well known that frequency-response masking (FRM) technique provides an attractive for realization digital filters very transition-bandwidths. In this paper, FRM design exploited applied to a novel cascaded bank realizing selective An...

Abstract With the expansion of epidemic, online multimedia teaching has become a common trend. The reasoning model evaluation is useful tool to infer result effects and predict tendency. However, ambiguity in linguistic-valued leads problems always context with uncertainty. To make better deal multiple multidimensional uncertainty environment, while considering both positive evidence negative at same time, this paper mainly focuses on linguistic truth-valued intuitionistic fuzzy layered...

Disputes may disturb construction projects and stakeholders, they cause tremendous losses that hinder the sustainable development of construction. Therefore, contractual governance is significant in as a crucial method dispute management. However, interrelation contract management has not been studied theoretically comprehensively. In this regard, paper aimed to propose framework for governance, including structures (GSs), mechanisms (GMs) an additional conceptual model, by using literature...

Generation Z looks to novel media, including podcasts, learn of the latest scientific innovations. This study compared use logical-scientific (LSC) narrative (NC) communication in science-based podcasts. Participants listened a podcast featuring LSC and NC while continuously rating their interest podcast. The section received higher average than section. different levels by science attitude level. Science podcasts should determine whether focus on or based target audiences.