Wavelet denoising with ICA for the segmentation of bio-acoustic sources in a noisy underwater environment

In this paper, comparative analysis of ICA algorithms has been done for separation of biological sounds in underwater noisy environment. Dolphin whistles, echolocation clicks and snapping shrimps are mixed to form underwater noisy environment. Various ICA algorithms like Fast-ICA, Infomax are applied on mixture of bio-signals. The efficiency of these algorithms has been compared using performance of separability. Signals available in open source and ICALAB based ICALAB simulation packages have been used to compare ICA algorithms.