Ophthalmic Images

OPHTHALMIC IMAGES

Ophthalmic images are common clinical exams to assess the normal operation of the eyes.
Angiography and eye fundus images record in a 2D-image the part of the eyes, abnormalities and blood vessels. On the other hand, optical coherence tomography and optical coherence tomography-angiography represent through a 3D volume the distribution of layers inside the retina and the vascularity inside them. Ophthalmic images can be saved and analyzed at a later time, and progression of diseases can be monitored over time.

 

A faster diagnosis of the disease grade may help to a proper clinical treatment, improving the patients quality of life. Some abnormalities inside the eyes can be detected by examining color photographs of the back of the eye, which is interpreted by a specialist ophthalmologist in retina. This process has some drawbacks such as being very time consuming and repetitive for clinical personnel, depending on the ophthalmologist’s experience, and its susceptibility to inter-observer variability. Moreover, the analysis of lots of images without any pathologies increases the work time, but decreases the time of analysis to others images with pathologies. Computer-aided systems (CADx) are an interesting alternative to tackling these problems CADx systems perform an automatic evaluation of the disease, they increase the number of patients diagnosed and reduce the time to detect diabetic eye diseases.

TEAM

Oscar Julián Perdomo Charry

Andrés Daniel Pérez Pérez

Melissa de la Pava Rodríguez

Victor Alfonso Arias Vanegas

PUBLICATIONS

Perdomo, O., Otálora, S., González, F. A., Meriaudeau, F., & Müller, H. (2018, April). OCT-NET: A convolutional network for automatic classification of normal and diabetic macular edema using sd-oct volumes. In Biomedical Imaging (ISBI 2018), 2018 IEEE 15th International Symposium on (pp. 1423-1426). IEEE. ISSN: 1945-8452. doi: 10.1109/ISBI.2018.8363839. (PDF).
Otálora, S., Perdomo, O., González, F., & Müller, H. (2017). Training deep convolutional neural networks with active learning for exudate classification in eye fundus images. In Intravascular Imaging and Computer Assisted Stenting, and Large-Scale Annotation of Biomedical Data and Expert Label Synthesis (pp. 146-154). Springer, Cham. doi: 10.1007/978-3-319-67534-3_16. (PDF).
Perdomo, O., Andrearczyk, V., Meriaudeau, F., Müller, H., & González, F. A. (2018). Glaucoma Diagnosis from Eye Fundus Images Based on Deep Morphometric Feature Estimation. In Computational Pathology and Ophthalmic Medical Image Analysis (pp. 319-327). Springer, Cham. doi: 10.1007/978-3-030-00949-6_38. (PDF).