We present a new efficient approach for detecting tumor of cancer in the human body based on a recently published discrete orthonormal space-frequency transform known as the DOST. We develop a frequency-domain implementation of the DOST in two dimensions for the case of dyadic frequency sampling. This manual detection may have chances of false detection. So a computerized method for cancer detection is needed. A feature selection algorithm based on statistical two-sample t-test method is used for the selection of significant coefficients from the high dimensional DOST(Discrete Orthonormal Stockwell Transform) coefficients. The selected significant coefficients are used as features for the classification of human body tumor as benign or malignant.