This volume contains 28 selected papers presented at IFTC 2024: 21st International Forum of Digital Multimedia Communication, held in Lingshui, Hainan, China, on November 28-29, 2024.

The 55 full papers included in this 2-volume set were carefully reviewed and selected from 146 submissions. They were organized in topical sections as follows:

CCIS 2441: Affective Computing, Graphics & Image Processing for Virtual Reality, Large Language Models, Multimedia Communication, Application of Deep Learning and Video Analysis.

CCIS 2442: Human and Interactive Media, Image Processing, Quality Assessment and Source Coding.

Inhalt

Spatio-Temporal Scene Graph Reasoning Networks for Emotion Recognition in User-Generated Videos.- Unsupervised 3D Face Reconstruction Method Based on ITV-Net.- Attention-Guided Semantic Segmentation Network for High Dimensional Multi-Scale Land Remote Sensing.- Reversible Data Hiding for Encrypted 3D Mesh Model Based on Optimal Grouping Strategy and Multiple-Bit Plane Prediction.- Semantic-Driven Free-View 3D Human Motion Video Composite.- RG-GS: Rasterization-Enhanced and Geometric-Guided Gaussian Splatting.- MoT: A Mixture of TriPlanes Framework for Frequency-Aware Dynamic Neural Radiance Fields.- High Quality 3D Gaussian Avatar Modeling.- Content Adaptive Light Field Representation Using Fourier Disparity Layers.- Research on Legal Question Answering System with Retrieval-Augmented Large Language Models.- Joint Source-Channel Coding with Large Language Model: A Vibrotactile Example.- Persona Extraction and Integration with Large Language Models Towards Personalized Dialogues.- MSBA: Adaptive Multi-Stream Data Transmission Method with Bandwidth Awareness for End-Cloud Systems.-Fabric Defect Detection Method Based on Unlabeled Compact Deep Learning.- Causal Imitation Learning-Based Navigation Algorithm for Drones.- Memory-Guided Hierarchical Feature Reconstruction for Multi-class Unsupervised Anomaly Detection.- A Method for Surface Defect Detection Based on Denoising and Self-Supervised Reconstruction.- Transmission Line Bolt Missing Detection Based on Improved YOLOv8 Network.- A Lightweight Infrared and Visible Image Fusion Method for Object Detection.- MSO-YOLO: Real-Time Pedestrian Detection Algorithm on Multi-Scale and Occlusion Situation.- Implicit Online Saddle Point Optimization.- MAFNet: Multi-Attention Fusion Network for Infrared Small Target
Detection.- Personalized Federated Meta-Learning Based on Gradient Clustering and Aggregation.- FMS-YOLO: Lightweight High-Altitude Work Safety Belt Detection.- Ink Animation Creation via Human-AI Collaboration.- A Meta-Space Architecture and Methods for Mobile Robot Inspection Digital Twin System.- Predict Pedestrian Flow in Open Street Environment.- VNNet: A Deep Learning-Based System for Video Visual Style Classification.