#$%^&*AU2020101581A420200917.pdf#####LYMPH NODE METASTASES DETECTION FROM CT IMAGES USING DEEP LEARNING ABSTRACT Due to advances in roller medical equipment and reduced costs in information storage, the absolute digital technology of the anatomical analysis of contaminated, infected tissue of computed tomography is becoming possible in recent decades. The current growth of online diagnosis includes online diagnoses, prompt access to oral history incidents, and faster communication with specialist surgeons. Detection of the existence of lymph node metastases is crucial to clinical pacing, diagnosis, and medication advice in females with cervical cancer. This proposal promotes a novel technique to detect the lymph node metastases of breast cancer using deep learning models to improve diagnostic accuracy and efficiency. The Deep Active Lesion Segmentation (DALS), a fully automatic segmentation mechanism without human intervention that optimizes a healthy non - linear extraction of features to obtain ultimately before Convolutional Neural Networks (CNNs). The convolutional network called inception v4 is proposed for generating the feature maps, and a deep neural network model like autoencoder is employed for the detection process. This invention offers essential treatment information of individual cancer patient's diagnoses and health care, and deep learning method provides excellent prediction capability and lymph node detection accuracy. 1 P a g eLYMPH NODE METASTASES DETECTION FROM CT IMAGES USING DEEP LEARNING Drawings CT scanner preprocessed automated segmented imag<； Inception v4 for feature maps image segmentation feature extraction meatssscore ~ ~ 2? 1'e~ Figure 1: Lymph node detection using DNN - ----------- fl - - ---Figure 2: Inception v4 Architecture for feature extraction 11 P a g e
