Saltar al contenido

2

Dedico esta entrada  a un interesante método de predicción de potencia de lente intraocular. Se ha desarrollado conjuntamente por el Dr Warren Hill, RBF (Radial Basis Function) Calculator Physician Team, Haag-Streit Suiza, y Mathworks.

Se basa en la red de función de base radial, un método desarrollado a finales de los años 80 como "neurona artificial". Es un método "hacia delante" en el que se introducen unos datos de entrada, se procesan asignándoles peso específico y se obtiene un valor de salida. Es esa capa oculta en la que se otorga peso específico la que se considera la neurona artificial.

Quizás la característica mas especial es que esta red neuronal artificial es capaz de auto aprendizaje y por lo tanto, según se vayan introduciendo casos de cirugías, las futuras predicciones serán mas ajustadas.

Actualmente este método está integrado en el software del biometro Lenstar pero puede accederse en la siguiente web: http://www.rbfcalculator.com/online/index.html

 

I dedicate this entry to an interesting prediction method of intraocular lens power. It has been jointly developed by Dr Warren Hill, RBF (Radial Basis Function) Team Physician Calculator, Haag-Streit Switzerland and Mathworks.

It is based on the network of radial basis function, a method developed in the late 80s as "artificial neuron". It is a method "forward", in which input data are entered, assigning specific weight are processed and an output value is obtained. The artificial neuron is this hidden layer in which the specific weight is considered.

Perhaps the most special feature is that this artificial neural network is capable of self learning and therefore as cases of surgeries are introduced, future predictions will be more tight.

Currently this method is integrated into the software of Lenstar biometer but can be accessed on the following website: http://www.rbfcalculator.com/online/index.html