As the demands of various network-dependent services such as Internet of things (IoT) applications, autonomous driving, and augmented and virtual reality (AR/VR) increase, the fifthgeneration (5G) network is expected to become a key communication technology. The latest video coding standard, versatile video coding (VVC), can contribute to providing high-quality services by achieving superior compression performance. In video coding, inter bi-prediction serves to improve the coding efficiency significantly by producing a precise fused prediction block. Although block-wise methods, such as bi-prediction with CU-level weight (BCW), are applied in VVC, it is still difficult for the linear fusion-based strategy to represent diverse pixel variations inside a block. In addition, a pixel-wise method called bi-directional optical flow (BDOF) has been proposed to refine bi-prediction block. However, the non-linear optical flow equation in BDOF mode is applied under assumptions, so this method is still unable to accurately compensate various kinds of bi-prediction blocks. In this paper, we propose an attention-based bi-prediction network (ABPN) to substitute for the whole existing bi-prediction methods. The proposed ABPN is designed to learn efficient representations of the fused features by utilizing an attention mechanism. Furthermore, the knowledge distillation (KD)- based approach is employed to compress the size of the proposed network while keeping comparable output as the large model. The proposed ABPN is integrated into the VTM-11.0 NNVC-1.0 standard reference software. When compared with VTM anchor, it is verified that the BD-rate reduction of the lightweighted ABPN can be up to 5.89% and 4.91% on Y component under random access (RA) and low delay B (LDB), respectively.

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC10007134PMC
http://dx.doi.org/10.3390/s23052631DOI Listing

Publication Analysis

Top Keywords

video coding
16
attention-based bi-prediction
8
bi-prediction network
8
versatile video
8
coding vvc
8
optical flow
8
proposed abpn
8
bi-prediction
6
network
5
coding
5

Similar Publications

Purpose: Head acceleration events (HAEs) are a growing concern in contact sports, prompting two rugby governing bodies to mandate instrumented mouthguards (iMGs). This has resulted in an influx of data imposing financial and time constraints. This study presents two computational methods that leverage a dataset of video-coded match events: cross-correlation synchronisation aligns iMG data to a video recording, by providing playback timestamps for each HAE, enabling analysts to locate them in video footage; and post-synchronisation event matching identifies the coded match event (e.

View Article and Find Full Text PDF

Education Research: Use by Neurologists of Microteaching and Microassessment Programs for Teaching, Learning, and Patient Care Needs: A Qualitative Study.

Neurol Educ

December 2024

From the Warren Alpert Medical School of Brown University (K.A.S.), Providence, RI; Memorial Sloan Kettering Cancer Center (A.M.M.), New York, NY; Department of Neurology (J.J.M.), Yale School of Medicine, New Haven, CT; Wake Forest University School of Medicine (K.W., S.-E.G., R.E.S.), Winston-Salem, NC; American Academy of Neurology (X.S., L.S., R.R., M.M., T.D.), Minneapolis, MN; and University of Michigan School of Medicine (Z.L.), Ann Arbor, MI.

Background And Objectives: Microlearning is the acquisition of knowledge or skills in small units, commonly delivered by digital technology. NeuroBytes (NB) and Question of the Day (QOD) are 2 microinstructional programs in neurology. NB programs are brief, video-based mini-courses on clinical topics (microteaching); QODs are daily multiple-choice questions (microassessment).

View Article and Find Full Text PDF

Objectives: Knowledgeable doctors are needed for timely assessment, diagnosis and management of lymphoedema. This qualitative study explored the thoughts and feelings of Australian interns (medical graduates in their first postgraduate year) towards (i) their understanding of the lymphatic system and lymphoedema, (ii) curricula pertaining to lymphoedema within their Australian medical degree and (iii) how they perceive that their understanding and medical training in lymphoedema influence their clinical practice.

Study Design And Methods: Qualitative semistructured interviews were conducted with interns employed within their first postgraduate year in Australia.

View Article and Find Full Text PDF

Background: Head-on-head impacts are a risk factor for concussion, which is a concern for sports. Computer vision frameworks may provide an automated process to identify head-on-head impacts, although this has not been applied or evaluated in rugby.

Methods: This study developed and evaluated a novel computer vision framework to automatically classify head-on-head and non-head-on-head impacts.

View Article and Find Full Text PDF

Objective: Presenting at academic conferences is an important means of disseminating research, networking, and building a professional reputation, but the quality of presentations at conferences is often suboptimal. This project describes the design, implementation, and evaluation of a presentation coaching program offered by an academic surgical society to presenters at its annual meeting.

Design: Oral presenters were paired with a coach and encouraged to meet independently, yet coaching was unstructured.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!