By Erik L. Ridley, AuntMinnie staff writer
November 8, 2010
Wednesday, December 1 | 12:15 p.m.-12:45 p.m. | LL-INS-WE3A | Lakeside Learning CenterIn this poster presentation, a University of Hong Kong research team will present its efforts to improve ultrasound image management by exploiting the images' textual information.
The group has had success applying image processing steps and optical character recognition technologies for use in ultrasound image management and data mining. With this approach, it's feasible to manage ultrasound files based on the labels already present on the images, even though it's not as accessible as the DICOM header information that's used in other images such as x-ray, CT, and MRI, said presenter Tao Chan.
"We believe such functions can become widely adopted when they are incorporated into standard PACS and further streamline [ultrasound] management," Chan said. "Besides the functions described in the paper and our previous work on anonymization, we have since then been developing tools for recording various [ultrasound] measurements based on the same principle of extracting text from the image content of DICOM files."
Last Updated hh 11/3/2010 2:38:54 PM