Abstract:
High rates of sickness and death from brain tumors make them a major public health
problem. This shows how important early discovery and correct classification are for
effective treatment. Traditional ways of diagnosing, like having doctors read Magnetic
Resonance Imaging (X-RAY) scans, take a lot of time, are prone to mistakes, and
depend on expert knowledge. Convolutional Neural Networks (CNNs) have come a
long way and show a lot of potential for automating the detection and classification of
brain tumors, with great accuracy and speed. Classical CNNs, on the other hand, have
a lot of problems, such as high computational costs, the ability to overfit, and problems
with processing big datasets quickly. Quantum Convolutional Neural Networks
(QCNNs) are a possible answer because they use quantum computing ideas like
superposition and entanglement to improve the ability to extract features and lower the
amount of work that needs to be done on the computer. This study used a large sample
of 4,599 X-RAY images to compare how well QCNNs and classical CNNs work at
finding brain tumors. The QCNN architecture includes quantum convolutional layers
and pooling layers that are meant to use quantum effects to make picture classification
more accurate. The classical CNN design, on the other hand, is made up of standard
convolutional layers, pooling layers, and fully connected layers that are optimized using
standard deep learning techniques. The X-RAY dataset was used to train and test both
models extensively, and accuracy, precision, recall, and the F1-score were used as
performance measures that were calculated and analyzed. QCNN did much better than
the standard CNN, scoring 97.82% on the test, which was a big improvement. Using
confusion matrices and Receiver Operating Characteristic curves, we were able to get
a lot of information about each model's diagnostic skills, strengths, and weaknesses.
