Skip to Main Content
The objective of this paper is to propose a framework in order to make the comparison of image processing algorithms effective. This framework was applied to three methods developed for automated regional wall motion scoring and they were compared to a reference scoring on a database of echocardiographic images (Evalechocard). Firstly, 200 annotated echocardiograms on hundred patients were used for the training stage; secondly the algorithms were blinded tested on 100 additional echocardiograms. Results obtained by the three methods are presented, using different metrics to compare them. This evaluation procedure enables a real progress in the assessment of each method and helps to understand its limits and its potentialities. Test results have shown the difficult cases and could be further used to improve the methods. Moreover the annotated database is now open to any research group who wants to test its own methods.