Image Analysis Project David Sinclair

1
Personal Medial Management
CVIeditor Computer Vision-based Image editor
ICON Image Content-based Organisation and
Navigation
Automated candidate region provision, selection
and colour change
The ICON system aims to enable users to easily
and effectively view, navigate and search
sizeable collections of images. This is achieved
through giving the system access to the semantic
content of an image by robust segmentation and
application of neural networks to recognise
everyday classes of stuff. A dedicated query
language, OQUEL, has been developed to allow near
natural language, text based image searching.
The CVIeditor is a novel region-based image
editor. It aims to automate many of the more
tedious tasks in image editing as well as
supporting more advanced user needs through
application of computer vision.
Video Analysis Annotation technology for
retrieval
Key-frame annotation
Archive
Video
Label motion events
OQUEL like search engine
GUI
Read closed caption
The video analysis part of the project is more
speculative and aims to develop an annotation
product for retrieval in commercial scale video
archives. An extension of OQUEL will be required
for specifying queries combining motion events,
text and visual content of key-frames.