Convolutional neural network and associated method for identifying basal cell carcinoma
AssigneeNational Cheng Kung University
InventorsGwo Giun Lee, Zheng-Han Yu, Shih-Yu Chen
Life of this patent
Estimating the term…
Abstract
A convolutional neural network (CNN) and associated method for identifying basal cell carcinoma are disclosed. The CNN comprises two convolution layers, two pooling layers and at least one fully-connected layer. The first convolution layer uses initial Gabor filters that model the kernel parameters setting in advance based on human professional knowledge. The method uses collagen fiber images for training images and converts doctors' knowledge to initiate the Gabor filters as featuring computerization. The invention provides better training performance in terms of training time consumption and training material overhead.
Claims7
A convolutional neural network, comprising layers in an order of:
- an input component receiving a batch of an image;
- a first convolution layer with initial Gabor filters that generates a first feature map;
- a first pooling layer reducing dimensions of the first feature maps;
- a second convolution layer with random kernels that generates a second feature maps;
- a second pooling layer reducing dimensions of the second feature maps;
- and
- a first fully-connected layer classifying the second feature maps into an output component, wherein
- each initial Gabor filter is defined into a Fourier transformation of Gabor-filter in an equation of:
- G(u,v;σx,σy,Ω,θ)=e−1/2((σx(u′−ω))2+(σyv′)2));
- u′=u cos θ+v sin θ;
- and
- v′=−u sin θ+v cos θ, where 0°≤θ<360°.
Dependencies are inferred from the claim text.