BMVA
British Machine Vision Association and Society for Pattern Recognition
http://www.bmva.org/meetings

Second Call for Participation

Visual Features in Medical Image Retrieval

One Day BMVA technical meeting in London, UK, on 25th April 2012
Chair: Dr. Xiaohong (Sharon) Gao, Middlesex University

Visual features have long played a vital role in image understanding and 
interpretation, especially in the medical domain. Therefore they have 
been at present increasingly being used to the task of indexing and 
retrieval. However, because of the gaps between low level features, 
e.g., colour, shape, and texture that computers are able to extract, and 
high level semantics from human’s interpretation, such as tumour or 
bleeding, visual feature based (or commonly known as content-based) 
image retrieval still remains in a research realm. On the other hand, 
the methods that are employed to extract and subsequently to interpret 
visual features vary appreciably, leading to very different groups of 
retrieved results. Particularly, the higher the dimension of an image 
(e.g., 2D, 3D, or 4D of a brain image) is, the more complexity towards 
the extraction of visual features can be. In general, the approaches 
that work perfectly on a set of 2D images will usually not function well 
for a 3D dataset, due to in part the ways that 2D algorithms constitute 
and in part the processing speed when the dimension is getting higher. 
In addition, due to the exponentially increase of medical images day by 
day, finding relevant data proves to be extremely difficult in a timely 
fashion, like finding a needle in a hay stack. Hence, visual feature 
based retrieval can help to a great extent for being complementary to 
the current text-based search.
This one day meeting aims at exchanging ideas in this field on the 
extraction of visual features with the focus on cementing the gap 
between low level visual features and high level semantics to ensure 
visual feature based retrieval is on course to be applied in a real 
clinical sector. The meeting will consist of tutorials, key note 
speeches and presentations with application areas include (but not 
limited to) medical images of 2D (e.g., x-ray, retinal), 3D or higher 
(e.g., CT, MR), and video images (e.g, ultrasound). The following 
topics, and any related, are encouraged to submit:

- Computational approaches to extraction of visual features
- Image semantics
- Visual feature interpretation and representation
- Image ontology
- Content-based image/video image retrieval
- Segmentation of salient visual features
- Image interpretation
- Medical image systems
- Medical data mining
- Image repository/database

Please submit an extended summary of about one A4-sized page (no longer 
than two pages) in length (PDF preferred by email attachment to Xiaohong 
(Sharon) Gao ([log in to unmask] ) by Friday 17th February 2012.

Dr Dimitrios Makris
Reader
Faculty of Science, Engineering and Computing
Kingston University, London

Tel: +44 20 8417 7082
Email: [log in to unmask] <mailto:[log in to unmask]>
Web: staffnet.kingston.ac.uk/~ku32195/ 
<http://staffnet.kingston.ac.uk/%7Eku32195/>

This email has been scanned for all viruses by the MessageLabs Email
Security System.