Registration of stereo and temporal images of the retina

• Published in: IEEE transactions on medical imaging Vol. 18; no. 5; pp. 404 - 418
• Main Authors: Ritter, N., Owens, R., Cooper, J., Eikelboom, R.H., Van Saarloos, P.P.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.05.1999; Institute of Electrical and Electronics Engineers
• Subjects: Algorithms; Biological and medical sciences; Fourier Analysis; Hierarchical systems; Humans; Image analysis; Image Processing, Computer-Assisted - methods; Image Processing, Computer-Assisted - statistics & numerical data; Image registration; Image sensors; Information theory; Interpolation; Investigative techniques, diagnostic techniques (general aspects); Magnetic Resonance Imaging - methods; Magnetic Resonance Imaging - statistics & numerical data; Magnetic sensors; Medical sciences; Multimodal sensors; Optic Nerve - anatomy & histology; Optic Nerve - diagnostic imaging; Optical films; Optical sensors; Otorhinolaryngology. Stomatology. Orbit; Photogrammetry - methods; Photogrammetry - statistics & numerical data; Positron emission tomography; Radiodiagnosis. Nmr imagery. Nmr spectrometry; Retina; Retina - anatomy & histology; Retina - diagnostic imaging; Sensor phenomena and characterization; Simulated annealing; Stereo vision; Time Factors; Tomography, Emission-Computed - methods; Tomography, Emission-Computed - statistics & numerical data; Tomography, X-Ray Computed - methods; Tomography, X-Ray Computed - statistics & numerical data; Radionuclide study; Human; Optic nerve; Radiodiagnosis; Similarity; Retina; Algorithm; Eye; Medical imagery; Measurement method; Hierarchical system; Information theory; Emission tomography
• ISSN: 0278-0062
• DOI: 10.1109/42.774168

The registration of retinal images is required to facilitate the study of the optic nerve head and the retina. The method the authors propose combines the use of mutual information as the similarity measure and simulated annealing as the search technique. It is robust toward large transformations between the images and significant changes in light intensity. By using a pyramid sampling approach combined with simulated reannealing the authors find that registration can be achieved to predetermined precision, subject to choice of interpolation and the constraint of time. The algorithm was tested on 49 pairs of stereo images and 48 pairs of temporal images with success.