Vasculatures are imperative structures in a human retina and their varied manifestations are often associated with abnormal state of many disorders. Automatic detection and analysis of these structures assist in diagnosis of many diseases such as diabetes, hypertension and arteriosclerosis. In this work, human retinal images are analysed using image processing techniques and transform based method. Normal and abnormal digital fundus images recorded under controlled protocol are employed for the study. The acquired images are subjected to Slantlet transform and the corresponding zero moments and statistical parameters from those features were derived for analysis. The derived parameters are correlated with vessel density index to identify vascular density and its variations. Results demonstrate that Slantlet transform is capable of extracting variations in vascular density in normal and abnormal images. Varying magnitudes of positive and negative peaks are observed for normal images whereas variations are found to be uniform for abnormal images. The values of second zero moment, skewness and kurtosis are found to correlate with vessel to vessel free area index. The correlation values were high for abnormal images than normal images. It appears that Slantlet transform based study carried out in this work seems to differentiate normal and pathological retina. As the analysis of retinal vessel features are important for pathological states related with retinal vein occlusion and tortuosity, these studies seems to be clinically relevant. In this paper, the objectives, methodology, results and the correlation analysis of all the derived parameters are presented in detail. © 2011 ISA.