The quality of a fish may be affected primarily by handling, processing and storage procedures from the catch to consumers. Retention time and storage temperature of post-harvested fish are key factors for sustaining the final quality of this product. This paper proposes an image processing method which is completely automatic, efficient and non-destructive for segmentation of tissues and prediction of freshness of the fish sample. The gill tissues of the fish sample are automatically segmented using a clustering based method and its features are strategically extracted in the wavelet transformation domain using Haar filter. First, second and third level decomposition in the wavelet domain is performed and the coefficients obtained at each level have been analyzed to predict the freshness of the fish sample. The experimental results indicate a monotonic variation pattern of the coefficients at the third level of decomposition and these coefficients gives an indication of the quality of the fish. This discriminatory variation in the image features with the duration of retention time provides a strategic framework for assessment of fish freshness. (C) 2015 Elsevier Ltd. All rights reserved.