History Made Every Day
The number and density of digital photos taken in cities is making it possible to automatically contextualize the photos and capture the continuous history across places, people, events, and objects of interest.
Precision-Recall Is Wrong for Multimedia
Editor's Note: I want to open a dialog about issues that are important to our field, and could change the direction of our work. I don't intend to write most of the columns, and I am soliciting ideas. Please send me your suggestions, either for issues ...
Visual Content Identification and Search
The past few years have witnessed rapid development and commercialization of visual content identification and search technologies that are concerned with identifying and searching visual content, particularly image and video content, by visual ...
Upcoming Events
This department presents publishing opportunities to multimedia researchers and announces pertinent upcoming conferences.
Visual Reranking: From Objectives to Strategies
A study of the development of visual reranking methods can facilitate an understanding of the field, offer a clearer view of what has been achieved, and help overcome emerging obstacles in this area.
Real-Time Video Copy-Location Detection in Large-Scale Repositories
By exploring the temporal relationships inherent in video, a probabilistic framework can help identify and locate copies of query videos.
Weighted Subspace Filtering and Ranking Algorithms for Video Concept Retrieval
The proposed framework, with weighted subspace filtering and ranking components, is the first attempt in multimedia research to apply multiple correspondence analysis to selected features while pruning data instances.
Naming People in News Videos with Label Propagation
A face-naming method that learns from labeled and unlabeled examples relies on iterative label propagation in a graph of connected faces or name–face pairs.
Discovering the Thematic Object in Commercial Videos
The thematic object in a commercial video is representative of its content. The authors propose a data-mining method for thematic object discovery in commercials by finding spatially collocated visual features.
Vocabulary Hierarchy Optimization and Transfer for Scalable Image Search
A vocabulary hierarchy optimization and transfer framework uses density-based metric learning to correct quantization biases, without supervision.
Music Generation with Markov Models
By using music written in a certain style as training data, parameters can be calculated for Markov chains and hidden Markov models to capture the musical style of the training data as mathematical models.
Mobile Visual Search: Architectures, Technologies, and the Emerging MPEG Standard
Editor's Note: Sophisticated visual-search and augmented-reality applications are beginning to emerge on mobile platforms. Such applications impose unique requirements on latency, processing, and bandwidth, and also demand robust recognition ...
New Products
This article features some of the latest advances in multimedia technology.