ACM Multimedia 2014
160 papers
ACM Multimedia 2014 featured 241 papers focusing on bridging the semantic gap in human activity recognition and decomposing appearance and motion within video models. Research directions included lifting 2D visual priors into 3D representations, exemplified by work on 3D fingertip detection and tracking. Another prominent theme involved integrating exogenous signals into social media dynamics, including the use of deep neural networks for psychological stress detection.
The edition also emphasized passive and unobtrusive sensing for continuous health and physiological monitoring. Practical applications presented included mobile video mashup systems, clothing retrieval based on local similarity, and interactive sketch recognition tools. Prolific contributors included Roger Zimmermann, Yang Yang, Nicu Sebe, and Shuicheng Yan. The conference also addressed automated video summarization and hosted the third workshop on geotagging and its applications. These works collectively reflected a trend toward combining social media analysis with longitudinal health data and improving 3D spatial understanding from 2D inputs.
Top Research Directions
Research directions most represented by ACM Multimedia 2014 papers.
Top Authors
Impactful Papers
Papers from ACM Multimedia 2014 that shaped research clusters.
Foundational
Reverbalize: A Crowdsourced Reverberation Controller
foundational
Mapping Semantic Descriptors to Audio Pr
Pedestrian Attribute Recognition At Far Distance
foundational
Leveraging Background Context for Visual
Multimodal Dynamic Networks for Gesture Recognition
foundational
Fusing Complementary Modalities for Robu
3D Human Activity Recognition with Reconfigurable Convolutional Neural Networks
foundational
Heterogeneous Spatiotemporal Feature Fus
A Dataset and Taxonomy for Urban Sound Research
foundational
Automated Diagnostic Classification from
Influential
Object Tracking using Reformative Transductive Learning with Sample Variational Correspondence
influential
Efficient Feature and Search Management
Menpo: A Comprehensive Platform for Parametric Image Alignment and Visual Deformable Models
influential
Designing Annotation Interfaces for Non-
libLDB: a library for extracting ultrafast and distinctive binary feature description
influential
Overcoming Platform Dependency in Specia
Pedestrian Attribute Recognition At Far Distance
influential
Biometric Recognition under Severe Resol
Object Tracking using Reformative Transductive Learning with Sample Variational Correspondence
influential
Robust Visual Tracking Under Appearance
Twitter-driven YouTube Views: Beyond Individual Influencers
influential
The Temporal Decay of Social Media Model