A sampling-based subband adaptive algorithm for speckle noise reduction of Optical Coherence Tomographic anterior chamber of eyeball images

This paper proposes a new wavelet thresholding algorithm to denoise Optical Coherence Tomographic(OCT) anterior chamber of eyeball image. The feature of this algorithm includes that (1) it is adaptive, every subband has its own threshold; (2) it is sampling-based so that it is not limited to any fixed noise pattern; (3) priori knowledge is needed, so the denoising process is more focused on individual images types; (4) a soft thresholding algorithm is applied to maintain the edge and remove the noise at the same time; (5)It is simplified which requires less calculation. Experimental result between the proposed algorithm and NormalShrink shows that the proposed algorithm has strong denoising performance and outperforms NormalShrink for OCT anterior chamber images.