부산대학교 도서관

In a nutshell, this abstract discusses a new model for brain tumour detection by using Convolutional Deep Learning (CDL) approaches such as Convolutional Neural Network (CNN), which is implemented in the PyTorch framework. The problems of brain tumor diagnosis are becoming more and more acute, requiring high-precision methods with maximum efficiency. The conventional practices often tell upon the man’s perception of medical imaging which is a tedious and subjective function. Unlike CNNs which use a data-driven approach and can Figure out the detection task with high precision, humans integrate this data to create intuitive judgements. Our technique is based on fast becoming an essential part of the deep learning community, PyTorch, a famous decision-making method. A set of complicated brain photos trained our model. The network architecture of the model was specifically intended to learn those complex patterns connecting with the tumour, and at the same time, it avoids confusing the normal ones as valid. With the help of strict experiments and evaluation, we show in no way uncertain results about the reliability of the method in tumour detections, even in the brain. The presented method was compared against other methods that exist and was demonstrated to be effective in terms of both accuracy and efficiency of the process. This work is an extension of the field-related computer-assisted diagnosis systems for brain tumours. This tool is vital for medical personnel in improving diagnostic accuracy and thereby enhancing the overall patient status.