Symbolic representation and visual querying of left ventricular image sequences

Area of Science:

• Cardiology
• Medical Imaging
• Computer Science

Background:

• Regional left ventricular function evaluation is crucial for diagnosing cardiac disorders.
• Current diagnostic methods rely on subjective analysis of image sequences, limiting reproducibility.
• Automated content-based retrieval can overcome these limitations for systematic analysis.

Purpose of the Study:

• To investigate the potential of content-based retrieval for analyzing cardiac image reports.
• To develop a system for automatic comparison of cineangiographic or echocardiographic images.
• To enable systematic and reproducible analysis of regional left ventricular function.

Main Methods:

• A system was developed for content-based retrieval from a database of cardiac images.
• Symbolic descriptions of geometrical and temporal properties of left ventricular contours were automatically derived.
• An iconic visual language was used for query expression and system interaction.

Main Results:

• The system automatically processes and interprets image data, associating symbolic descriptions with reports.
• Queries expressed via an iconic visual language are automatically interpreted and compared against database descriptions.
• Medical users can effectively search for motion and shape abnormalities on a regional basis.

Conclusions:

• Content-based retrieval offers a systematic and reproducible approach to analyzing regional left ventricular function.
• The developed system facilitates both prospective and retrospective diagnosis by enabling efficient searching of cardiac image reports.
• This technology has the potential to significantly enhance the diagnostic capabilities in cardiology.