ABSTRACT
A novel splat feature classification method for retinal haemorrhage detection
is presented in this project. The retinal colour images are partitioned into non
overlapping segments covering the entire image. Each segment contains pixels with
similar color and spatial location. The set of features are extracted from each splat to
describe its characteristics relative to its surroundings, employing responses from a
variety of filter bank, interactions with neighbouring splats, shape and texture
information. The optimal subset of splat features is selected by a filter approach. The
Gaussian filter is also used for feature extraction. The splat features are extracted by
using Gaussian filter bank technique. The pre-processing is done by the Histogram
equalization technique. For segmentation, watershed transform is used to detect the
retinal haemorrhage. In this project, the receiver operating characteristics curve at the
splat level and the image level are being taken into consideration. Watershed
algorithm is used to segment the images and the SVM classifiers are used to classify
the segment of images. The MATLAB program was written to detect the retinal
hemorrhages in a diabetic patients.
Need projects in bio-medical contact us
