Summarization of echo-Doppler videos for computer-aided diagnosis

Summarization of echo-Doppler videos allows reducing diagnosis time, improving comparison between videos and making a more efficient storing of them. This paper aims at providing a solution to the static summarization of echo-Doppler videos. A static summary of a video is a collection of keyframes together with a description of them. The selection of keyframes discussed in this paper is based on the analysis of properties of red blobs resulting from Power Doppler technique. The properties of red blobs are extracted by a robust thresholding algorithm in the HSL colour model and via a connected-component labelling algorithm. Keyframes extracted from the echo-Doppler video satisfy specific properties for the red blobs. The work is still in progress and we are now collecting data for the construction of a decision support system to help doctors in their diagnoses. First results are encouraging and future works will bring to an interesting computer-aided system.