本书汇集了张勤及其团队在智能音视频处理领域近年来的研究成果及前瞻性思考,涵盖情智信息、媒介音频、视觉处理和人工智能四个部分。情智信息部分探讨了情感计算在智能系统中的应用,包括EEG 情感识别、音乐情感质量评估等多个方面;媒介音频部分展现了如何通过智能技术,在音乐创作、音频处理等领域开辟新的可能性;视觉处理部分通过 3D 重建、人体姿态估计等领域的重要突破,展示了智能技术在空间感知和虚拟现实等方面的应用潜力;人工智能部分聚焦神经网络模型的创新与优化,指出了人工智能在模型理解和优化中的新方向。
第一部分 情智信息
情智信息的建模与应用
Interaction Between Dynamic Aff ection and Arithmetic Cognitive Ability: a Practical
Investigation with EEG Measurement
SSTM-IS: Simplifi ed STM Method Based on Instance Selection for Real-Time EEG
Emotion Recognition
Multi-Source Information-Shared Domain Adaptation for EEG Emotion Recognition
Emotional Quality Evaluation for Generated Music Based on Emotion Recognition
Model
第二部分 媒介音频
Design of Linear-Phase Nonsubsampled Nonuniform Directional Filter Bank with
Arbitrary Directional Partitioning
Multi-Source Separation Using over Iterative Empirical Mode Decomposition
A Two-Stage Complex Network Using Cycle-Consistent Generative Adversarial Networks for Speech Enhancement
Analysis of Music Rhythm Based on Bayesian Theory
Learning to Generate Emotional Music Correlated with Music Structure Features
Visually Aligned Sound Generation via Sound-Producing Motion Parsing
MovieREP: a New Movie Reproduction Framework for Film Soundtrack
第三部分 视觉处理
Distributed Markov Chain Monte Carlo Kernel Based Particle Filtering for Object Tracking
Use Hierarchical Genetic Particle Filter to Figure Articulated Human Tracking
Human Action Recognition Using Multi-Velocity STIPs and Motion Energy Orientation Histogram
Semantic Based Autoencoder-Attention 3D Reconstruction Network
Flexible Light Field Angular Superresolution via a Deep Coarse-to-Fine Framework
Cross-Domain Feature Similarity Guided Blind Image Quality Assessment
A Dataset and Benchmark for 3D Scene Plausibility Assessment
第四部分 人工智能
Interpretability Diversity for Decision-Tree-Initialized Dendritic Neuron Model Ensemble
Pruning of Dendritic Neuron Model with Signifi cance Constraints for Classifi cation
A General Paradigm of Knowledge-Driven and Data-Driven Fusion