The shape and appearance of the optic nerve head region are sensitive to changes associated with glaucoma and diabetes that may be otherwise asymptomatic. The changes can be diagnostic of the diseases, and tracking of the changes in sequential images can be used to assess treatment and the progress of the illness. At present, change detection and tracking are performed manually, which can be a cause of poor repeatability. We are concerned with developing automated techniques of generating quantitative descriptions of the retinal images that might be used in diagnosis and assessment. In this paper, we investigate the use of images that have been collected and stored remotely, as this will replicate capture and automated processing by outreach clinics. Normal and abnormal images were collected from a range of sources, to simulate the mass screening process. The images were processed using simple signal-processing methods and divided into two groups. Using a chi-squared test, the separation of normal and abnormal images using this test was found to be highly significant (p < 0.05, n = 60).

Many domestic appliances and much office equipment is controlled using a keypad and a small digital display. Programming such devices is problematical for the blind and visually handicapped. In this paper, we describe a device that may be used to read the displays on these devices. The device is designed to accept a description of the display being read, which specifies the types and locations of elements of the display. Images are captured using a handheld webcam. Images are processed to remove the distortions due to the camera orientation. The elements of the screen are interpreted and a suitable audio output is generated. In suitably illuminated scenes, the display data is interpreted correctly in approximately 90% of the cases investigated

Binary Robust Independent Elementary Feature (BRIEF) has been successfully used as an efficient feature point descriptor. It uses can be found in many computer vision application such object recognition and camera localization. This is due mainly to its discriminating ability and computational simplicity as well as invariant to illumination variations. In this paper we use a slightly modify BRIEF descriptor as a method to measure texture and apply them in texture segmentation application. Experiments are conducted on a standard test suite of test images and shows that BRIEF texture features is better and in some cases compare favorably with other established texture measures.