Investigation of multiwavelets and set partitioning algorithm on mammogram images
Uthandipalayam Subramaniyam Ragupathy, Angamuthu Tamilarasi · International Journal of Modelling Identification and Control · 2010
Breast cancer has the highest death incidence rates among women, ranking next to lung cancer. Mammographic screening is the best tool to detect cancerous lesions. Mammographic image of high resolution occupies large size, hence compressing it by preserving the information is necessary for easy storage and transmission. Advances in wavelet transforms and quantisation are capable of surpassing the existing image compression standards like Joint Photographic Experts Group (JPEG) algorithm. Wavelet-based Set Partitioning In Hierarchical Trees (SPIHT) algorithm gives better compression. Wavelet transforms require filters that combine desirable properties like orthogonality and symmetry, but they cannot possess these properties simultaneously. Multiwavelet offers these desirable transform features. But there are some limitations with SPIHT algorithm for multiwavelets coefficients. This paper used a method called coefficient shuffling for encoding the multiwavelet decomposed images by shuffling the coefficients as suitable for SPIHT algorithm, and is investigated on mammographic database which gives better compression performance.