The segmentation of tumour from a cancer
MRI images in image processing is classic research area
of interest and a tedious task. Manually segmenting the
MRI images is very time consuming and liable to errors.
Many researchers have done investigation using deep
neural network in segmenting the oral MRI images as
they poses higher performance in segmenting the oral
cancer images automatically. Owing to their gradient
dissemination and complexity issues, the CNN takes
more time and excess computational power in training
the images. Our aim is build an automated technique for
the segmentation of oral cancer images using Residual
learning networks (ResNet) to render the complications
of gradient dissemination caused by CNN. ResNet attains
higher accuracy and trains the images faster compared
to CNN. To accomplish this, ResNet counts a skip
connection parallel to convolution neural network layers.
The verification accuracy of the proposed technique has
been carried out on oral cancer (lip and tongue) images
dataset. The results of proposed technique shows a better
accuracy, dice co-efficient, specificity and precision of
0.92, 0.95, 0.94, 0.96 respectively and computational time
of 63 mins.
Keywords :
Oral cancer, Segmentation, DNN, ResNet.