Multimodal aspect-based sentiment analysis (MABSA) is an essential task in the field of analysis, which still confronts several critical challenges. The first challenge how to effectively capture key information within both image and text features enhance recognition understanding complex expressions. second achieve cross-modal alignment multi-granularity features. third narrow semantic gap between modality through effective feature fusion. To address these issues, a framework that leverages...

In this paper, we propose an extension to current TCP protocol named concurrent (cTCP), which can improve throughput of connections through balancing traffic load on multiple end-to-end paths. We keep the single-sliding-window architecture untouched for compatibility reasons. The two core components design are a packet scheduling switch that balances among paths, and duplicated acknowledgement classifier handles reordering. implemented our in FreeBSD kernel compared performance cTCP with...

Abstract The fusion of digital media art with film and television animation production ushers in a new artistic technological synergy era. This form art, merging visual, sound, innovations, faces unparalleled growth opportunities significant challenges. paper investigates the impact on animation, emphasizing its influence creative expression processes. Through case study analysis, we demonstrate efficiency gains aesthetic enhancements afforded by media, including 70% reduction rendering time...

In this paper, we propose an extension to current TCP protocol named concurrent (cTCP), which can improve throughput of connections through balancing traffic load on multiple end-to-end paths. We keep the single-sliding-window architecture untouched for compatibility reasons. The two core components design are a packet scheduling switch that balances among paths, and duplicated acknowledgement classifier handles reordering. implemented our in FreeBSD kernel compared performance cTCP with...

With IT and video technologies evolving rapidly, services with user interaction emerge, bring about attracting experiences. The growing bandwidth latency demands call for advanced transmission techniques which utilize wired, wireless, 5G cellular network, cooperate unicast, multicast, broadcast. Free-viewpoint video, as one of the interactive applications, can provide users immersion real world scenes, shows potential to be an essential application in future digital life. However, such is...

5G and edge computing have brought great changes to video industry. Interactive is becoming an emerging application form of multimedia service, which provides attraction beyond typical scenarios like cloud gaming remote virtual reality (VR), puts forward challenges in resource capacity, response latency, function flexibility its service system. In this paper, we propose elastic system architecture with low latency features accommodate generic interactive applications on near user edges. To...

Videos transmitted on the Internet need multiple versions due to variety of needs from users. Video providers transcode an original video into different formats, with bitrates and resolutions satisfy those needs. To meet such demands, we design implement a novel lightweight containerized media cloud for transcoding service in this paper. We show high efficiency, robustness, scalability portability our system.

Based on the improved Watson's model, an adaptive spread-transform dither modulation (ASTDM) algorithm is proposed for efficient watermarking to resist amplitude scaling and coefficient re-quantization attacks. Firstly, luminance masking modified be consistent with without affecting original parameters. Then, by incorporating model STDM algorithm, a maximum labeling strength appropriate quantization step-size shrinking local values of host image are achieved. Experiments demonstrate that...

As an important application form of immersive multimedia services, free-viewpoint video (FVV) enables users with great experience by strong interaction. However, the computational complexity virtual view synthesis algorithms poses a significant challenge to real-time performance FVV system. Furthermore, individuality user interaction makes it difficult serve multiple simultaneously for system conventional architecture. In this paper, we novelly introduce CNN-based interpolation algorithm...

Concurrent transfer is a feasible solution to improve the efficiency and reliability of data communication. In this study, we developed concurrent transmission control protocol (cTCP), which based on TCP but allows multi-path transfer. Our cTCP effectively minimizes impact packet reordering problem under experimental result confirms that can achieve theoretical optimal or near performance

One of the principal contradictions these days in field video i s lying between booming demand for evaluating streaming quality and low precision Quality Experience prediction results. In this paper, we propose Convolutional Neural Network Gate Recurrent Unit (CGNN)-QoE, a deep learning QoE model, that can predict overall continuous scores services accurately real time. We further implement state-of-the-art models on basis their works compare with our method six public available datasets....

As short video industry grows up, quality assessment of user generated videos has become a hot issue. Existing no reference methods are not suitable for this type application scenario since they aimed at synthetic videos. In paper, we propose novel deep blind model according to content variety and temporal memory effect. Content-aware features frames extracted through neural network, patch-based method is adopted obtain frame score. Moreover, memory-based pooling considering effect predict...

There are many resolution enhancement methods to produce 4K-UHD videos including interpolation and deep learning based Super-Resolution, but lack of details, aliasing unnatural transition still inevitable affect viewer's immersive experience. Therefore, it is necessary detect authenticity native videos. In this paper, we propose a new detection method for The on extracting feature from the DCT frequency domain integrating in all frames. Our can determine as well resolution, no matter which...

In this paper, we present methods to enhance the fault tolerance of large-scale optical switches through innovations in architecture and control logic design. A switch is constructed from a network 2/spl times/2 elements (SEs). Classic architectures, such as Benes, are not designed with mind. There three major contributions paper: 1) developed an analytical method, referred probability accumulation calculate average connection blocking faulty network; 2) provided fault-aware routing...

Portable smart devices have become popular in recent years. These also produce a large number of data. In wireless environment, it is important for portable to keep reliability data transmission. While difficult stable environment. Instability path may cause unreachability between the server and devices. this state, some real-time services unavailable order overcome problem, we design novel fault tolerating system The architecture allows applications use multiple connections access...

Remote sensing of ocean surface pressure from space is very important, and differential absorption LIDAR radar are only two kinds remote instruments with this potential. The works the integral path mode spacecraft in 400 km low-Earth orbit. optical depth oxygen A-band measured, then obtained using a circle-iterative calculation. Performance evaluation model was based on feasibility to advanced system parameters instrument, whilst weak echo pulse energy at yielded random errors measurement....