Deep learning for diabetes disease

We are interested on giving a kind of diabetes disease diagnosis from eye images processing and analysis. Particularly, in this part of research we deal with object identification from Fundus photography. We analyse several kind of fundus images traying to identify:

  • Optic nerve and
  • Macula

The optic nerve, also known as cranial nerve II, is a paired nerve that transmits visual information from the retina to the brain.

The macula is the part of the eye responsible for central high acuity vision.

Detection of the macula and optic nerve is an important task in retinal image processing as a landmark for subsequent disease assessment, such as for age-related macula degeneration.

With that aim we work with Convolutional neural network combined with vision algorithms. Below are some of our results.

Optic nerve identification

Automatic detection of the optic nerve in retinal fundus images. The effectiveness of these results is around 95% of the test group of images. The green square is the result of the algorithm for optic nerve identification.

Macula identification

Automatic detection of the macula in retinal fondus images. We have tested our algorithm in several kind of images in illumination, color and size are considerably variable.

Students

Tec. Gesem Eliab Gudiño Mejía: Macula and optoic nerve identification by Convolutional neural network; Universidad Tecnológica de León y Centro de Investigaciones en Óptica, A. C; 2017-2018.

Ing. Michel Alain González Pascual; Tecnológico de Estudios Superiores de Ecatepec y Centro de Investigaciones en Óptica, A. C 2016-2017.

Tec. Rodolfo Isaac Verdín Monzón; Universidad Tecnológica de León y Centro de Investigaciones en Óptica, A. C 2016-2017.