Wyniki wyszukiwania - BazTech

Ograniczanie wyników

2 Machine Graphics and Vision

2 1998

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: knowledge-guided image analysis

Sortuj według:

Ogranicz wyniki do:

Image processing and knowledge-guided data analysis for medical diagnostics support

Belikova T. P., Stenina I. I., Yashunskaya N. I.

Machine Graphics and Vision

1998

Vol. 7, No. 4

841-858

A series of methods has been developed to support medical image analysis. Optimal filtering was applied for better imaging of informative features (diagnistically important details and structures). several models of image background were tested to desing an optimal filter, which is the best by expert criteria. Descriptions of processed images were made in terms of emphasized informative features, collected in the database, and used to find a decision rule, which presented an effective solution of medical tasks in easily interpretable form. Developed methods allowed to explicit some elements of human visual task structure into several simpler subtasks. It shows the approaches to effective learning the physicians and to automated computing some informative features directly in the image to realize expert-independet part of image analysis and description. The methods were tested in the task of early peripheral lung cancer diagnisis and gave an essential improvement of diagnostic accuracy for the physicians of different qualifications.

Image interpretation based on image processing and knowledge-guided data analysis

Belikova T. P., Stenina I. I., Yashunskaya N. I.

Machine Graphics and Vision

1998

Vol. 7, No. 1/2

187-196

A series of methods has been developed to promote medical image analysis and interpretation in complex diagnostic task. Image processing methods wher used for better representation of informative features for the expert's analysis. Then the expert's descriptions of processed image where colected in the database to find discriminative features and create classification decision rules. Created threshold rules were interpreted by physicans as a syndrome- like construction conventional for medicine. We proposed some statistics that were helpful for measuring some discrimenative features directly in the image ro realze unbiased expert-independed part of image descriptions. Developed methods were applied for early peripheral lung cancer diagnosis and helped to improve the efficacyb of image interpretation for phisicians of different qualifications.