Abstract: Intensity degradation is a common problem in noise filtering for fluorescein angiogram sequences. In this paper a new filter model, extended adaptive weighted averaging (EAWA) filter is proposed for the purpose of reducing noise of a fluorescein angiogram sequence by adopting the method of adaptive weighted averaging and the idea of order statistic filter. The EAWA filter is a flexible filter model, by which even a particular filter can be designed via defining pixel of interest (POI) and the sub-support of the filter accord ing to some specified requirements. Based on the EAWA filter model, a fluorescei n-angiogram-oriented filter, HIP filter is introduced to solve the problem of high-intensity preservation on filtering noise. The results of the contrast experiment show that the HIP filter has the features of frame-averaging, which can achieve noise suppression efficiently, and the features of high-intensity preservation for bright pixels, which can keep the intensities of bright pixels from degrading via distributing the weights adaptively according to the inte nsity of each pixel.

Key words: fluorescein angiogram, image sequence, noise filtering, extended adaptive weighted averaging filter, adaptive weighted averaging filter, pixel of interest