Stephen Taylor, Computational Biology Research Group, Oxford
Biological imaging is generating an unprecedented amount of data. New, sophisticated imaging techniques yield large, multidimensional data sets that need to be viewed, analysed, annotated, queried and shared. As high throughput informatics tools have enabled large-scale dna sequence analysis, analysing large numbers of images requires image informatics tools to be improved. Image analysis techniques usually require visual inspection and verification of the analysis results, but access to raw and derived results of many images rapidly becomes a challenge. Even collections of manually annotated images may be difficult to comprehend since users frequently want to assess features that are not accurately quantified in the metadata.
A key aim of this research is to facilitate mining of complex image data sets, allowing users to quickly filter out false positive leads and combine querying with the human brain's innate pattern recognition ability. Zegami allows visualisation and seamless querying of associated metadata across collections of thousands of images and movies in any web enabled device.
The talk will show examples from microscopy but also will highlight other applicable domains including high throughput screening, sequence analysis and document management.