Purpose The purpose of this paper is to build a mobile visual search service system for the protection Dunhuang cultural heritage in smart library. A novel model murals proposed help users acquire rich knowledge and services conveniently. Design/methodology/approach First, local global features images are extracted, dictionary generated by k -means clustering. Second, based on bag-of-words (BOW) multiple semantic associations constructed. Third, library designed cloud environment....

Purpose The deployment of vaccines is the primary task in curbing COVID-19 pandemic. purpose this paper to understand public’s opinions on and then design effective interventions promote vaccination coverage. Design/methodology/approach This proposes a research framework based spatiotemporal perspective analyse public opinion evolution towards vaccine China. first obtains data through crawler tools. Then, with help mining technologies, such as emotion computing topic extraction,...

In this study, composite plates of 6061/TA1 were successfully manufactured using additive friction stir deposition (AFSD). The impact preheating temperatures (room temperature, 100 °C, 200 °C) on the interfacial microstructure and interface mechanical properties at various zones was studied. results showed that as temperature increased or when deposit zone shifted from boundary to center, diffusion width Al Ti increased, accompanied by an increase in bonding shear strength. Moreover, sample...

Purpose Aspect category-based sentiment analysis (ACSA) has been widely used in consumer preference mining and marketing strategy formulation. However, existing studies ignore the variability features intrinsic correlation among diverse aspect categories ACSA tasks. To address these problems, this paper aims to propose a novel integrated framework. Design/methodology/approach The framework consists of three modules: text feature extraction fusion, adaptive selection category-aware decision...

Purpose Previous research mainly uses graph neural networks on syntactic dependency graphs, often neglecting emotional cues in sarcasm detection and failing to integrate image features for multimodal information effectively. To address these limitations, this study proposes a novel model based the directed isomorphism network with sentiment enhancement fusion (DGIN-SE-MF). Design/methodology/approach The approach extracts text through vision transformer BERT, respectively. deeply extracted...

Abnormal behavior has become the leading cause of injuries among elderly in modern society. Elderly anomaly is a widespread concern both academic and industrial fields. Detection abnormal behaviors using deep learning techniques been pervasively investigated due to flourishing powerful edge computing decreasing costs consumer electronics, such as commercial vision sensors. This paper intentionally designs an innovative architecture detect typical behaviors: falls tumbles, aggression...

In this study, the ball-on-disk sliding wear and tribocorrosion behavior in H2SO4 HCl solution of NiCoCrMoCu alloys with carbon additions 0.2, 1, 1.5, 2 wt.% Al2O3 ball as a counterpart was investigated systematically. Obvious antagonistic effects were found after both aqueous solutions. Compared dry conditions, lubrication effect significantly reduces rate alloy, reduction more obvious than that HCl. The 0.2C 1C decrease load rate, while those 1.5C 2C increase. (coefficient friction) COF...

Purpose The paper aims to construct a spatiotemporal situational awareness framework sense the evolutionary situation of public opinion in social media, thus assisting relevant departments formulating control measures for specific time and space contexts. Design/methodology/approach comprises element extraction, understanding projection. In data on COVID-19 vaccine, including tags text contents, is extracted. understanding, bidirectional encoder representation from transformers – latent...

The prosperity of e-commerce has changed the whole outlook traditional trading behavior. More and more people are willing to conduct Internet shopping. However, massive product information provided by Merchants causes problem overload this will reduces customer’s satisfaction interests. To overcome problem, a recommender system based on web mining is proposed in paper. utilizes techniques trace shopping behavior learn his/her up-to-date preferences adaptively. experiments have been conducted...

The uniformity and microstructure of the copper deposition in high aspect ratio plated through holes (penetrating holes) are crucial for performance printed circuit board. We systematically investigated effects reverse pulse parameters period (PPR) plating on deposition, including frequency, duty cycle current density. Cu behavior (throwing power) its crystallographic characteristics, grain size, orientation, boundary, were characterized by means field-emission scanning electron microscopy...

Purpose The outbreak of COVID-19 has become a major public health emergency worldwide. How to effectively guide opinion and implement precise prevention control is hot topic in current research. Mining the spatiotemporal coupling between online offline epidemics can provide decision support for management future emergencies. Design/methodology/approach This study focuses on analyzing relationship epidemic. First, based Weibo information confirmed case information, field framework constructed...

Purpose Dunhuang murals are rich in cultural and artistic value. The purpose of this paper is to construct a novel mobile visual search (MVS) framework for murals, enabling users efficiently similar, relevant diversified images. Design/methodology/approach convolutional neural network (CNN) model fine-tuned the data set murals. Image features extracted through CNN model, similarities between different candidate images query image calculated by dot product. Then, sorted similarity, semantic...

Murals unearthed in China have outstanding regional characteristics and one of the largest period spans scale variety. To explore visual distinguishability topic autocorrelation murals from spatial perspective, multiple classification models are employed to classify through features. Then, k-means is mine topics, they analysed intensities (TIs), Moran’s Index (MI) concentration degrees (STCDs). In addition, distribution evolution summarised revealed dimension. From a it verifies features...

This paper presents a monocular camera (MC) and inertial measurement unit (IMU) integrated approach for indoor position estimation. Unlike the traditional estimation methods, we fix downward to floor collect successive frames where textures are orderly distributed feature points robustly detected, rather than using forward oriented in sampling unknown disordered scenes with pre-determined frame rate auto-focus metric scale. Meanwhile, adopts constant scale adaptive determined by IMU data....

Purpose Twitter fake accounts refer to bot created by third-party organizations influence public opinion, commercial propaganda or impersonate others. The effective identification of is conducive accurately judge the disseminated information for public. However, in actual account identification, it expensive and inefficient manually label accounts, labeled data are usually unbalanced classes. To this end, authors propose a novel framework solve these problems. Design/methodology/approach In...