Loading...

A Ring Selection Platform for Treatment of Keratoconus

Khademi Mofrad, Amir Hossein | 2022

50 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56345 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Asghari, Mohsen
  7. Abstract:
  8. Corneal keratoconus is one of the common eye diseases that usually occurs in the teenage years or the beginning of the third decade of life. In this disease, the cornea thins and loses its original shape and becomes conical. To treat this disease, ophthalmologists use different methods such as using glasses, contact lenses (hard and soft), corneal transplantation, and also corneal cross-linking, each of these methods has limitations, on the other hand, considering that The disease progresses and its severity increases, Ophthalmologists use corneal rings to treat this disease in more advanced stages, implanting a suitable ring in the patient's cornea both increases the strength of the patient's cornea and causes the cornea to return to its original state, on the other hand, considering that ophthalmologists still do not have enough information about They do not have the right ring geometry for every patient with keratoconus, and they only use their own experiences, so by using the simulation of intracorneal ring implantation, very useful information can be provided to ophthalmologists. In this research, at first, the cornea model of a patient was prepared using ophthalmology data, and after allocating its properties according to previous research, the simulation of intra-corneal ring implantation with the variable consideration of ring radius, angle, planting height and planting area factors and 288 simulations have been done, after that with the help of genetic algorithm and extraction of Zernik coefficients, we obtain the parameters related to the quality of vision for each simulation mode. Two algorithms of random forest and deep learning of the data are trained and a platform is created that can predict the outcome of the operation for the eye of the same patient by changing each of the factors. After checking the prediction accuracy of the trained data on a specific model, we create a database by applying variables related to the location and geometric shape of the hump and repeating the previous simulations. Through the mentioned algorithms, it is possible to predict the quality of vision for new models with different hump and protrusion conditions, thus, by predicting the outcome of the operation in different conditions, doctors can choose the best possible conditions for the surgery
  9. Keywords:
  10. Keratocanic Corneas ; Machine Learning ; Finite Element Method ; Ring Implantation ; Random Forest Algorithm ; Deep Learning

 Digital Object List

 Bookmark

No TOC