Recent news
The MICC at the MatchMaking ICT Florence
ToscanaIN and MICC Workshop "The multisensorial design" at CNA Next Week
A coarse-to-fine approach for fast deformable object detection
Inauguration of Urban Innovation Park of the City of Florence
MICC cluster presentation
On automatic reading of handwriting (What the writing hand could tell to the reading eye)
Vehicles that Learn by Observing How We Drive: Multidisciplinary Explorations in Human-Centered Driver Assistance
Developmental Agents for Vision
Last months lectures
Scientists and researchers from around the world have regularly lectures at our Center.
On automatic reading of handwriting (What the writing hand could tell to the reading eye)
by
Vehicles that Learn by Observing How We Drive: Multidisciplinary Explorations in Human-Centered Driver Assistance
Analysis and development of a learning application for people with autism spectrum disorder (ASD) according to the Applied Behavioral Analysis (ABA) method
loading..
loading..
- Lecturer: Andrew Bagdanov received his PhD in computer science in 2004 from the University of Amsterdam. He held a postdoctoral position for three years at the University of Florence, where his research focused on active camera control, visual tracking and detection, and retrieval of trademarks in sports videos. After Florence, Dr. Bagdanov held a senior development position at the FAO of the United Nations in Rome, where he worked on developing large-scale, multilingual ontologies for cross-language retrieval of agricultural information in over twenty languages. Andy is currently a senior research fellow at the Computer Vision Center, Barcelona, where he continues to work in the fields of active camera control, multiview geometry and semantic image analysis and retrieval. His research spans a broad spectrum of computer vision, image processing and machine learning. He has published over fifty scientific articles in peer-reviewed, international journals and conference proceedings.
- Custom: Andrew Bagdanov
The harmony potential: fusing local and global information for semantic image segmentation
Semantic image segmentation is the process of assigning semantically relevant labels to all pixels in an image. Hierarchical Conditional Random Fields (HCRFs) are a popular and successful approach this problem. One reason for their popularity is their ability to incorporate contextual information at different scales. However, existing HCRF models do not allow multiple labels to be assigned to individual nodes. At higher scales in the image, this results in an oversimplified model, since multiple classes can be reasonable expected to appear within a single region. This simplified model especially limits the impact that observations at larger scales may have on the CRF model. Furthermore, neglecting the information at larger scales is undesirable since class-label estimates based on these scales are more reliable than at smaller, noisier scales.
The harmony potential: fusing local and global information for semantic image segmentation
In this talk I will discuss a new potential function, the harmony potential, for defining HCRF models of semantic image segmentation. The harmony potential can encode any possible combination of class labels at the global level, enabling it to make better informed, fine discriminations at the low levels. This representational capacity of the harmony potential is also its primary weakness as the optimization over all possible labels quickly becomes intractable for more than a few classes. To address this, we show how the harmony potential model admits an effective sampling strategy that renders tractable the underlying optimization problem. Results show that our approach obtains state-of-the-art results on two challenging datasets: Pascal VOC 2009 and MSRC-21. The approach described in this talk additionally won six gold medals in the Pascal VOC 2009 Segmentation Challenge.