Automating cancer diagnosis using advanced deep learning techniques for multi-cancer image classification.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Cancer is a leading cause of death globally, necessitating early detection.
• Traditional methods are often invasive and time-consuming.
• There is a need for efficient and accurate automated cancer detection solutions.

Purpose of the Study:

• To evaluate deep learning models for automated cancer detection.
• To compare the performance of various Convolutional Neural Networks (CNNs).
• To identify the most effective AI model for multi-cancer image analysis.

Main Methods:

• Utilized deep learning models including DenseNet121, DenseNet201, Xception, InceptionV3, MobileNetV2, NASNetLarge, NASNetMobile, InceptionResNetV2, VGG19, and ResNet152V2.
• Applied image segmentation and contour feature extraction (perimeter, area, epsilon).
• Evaluated models on image datasets for seven cancer types: brain, oral, breast, kidney, Acute Lymphocytic Leukemia, lung and colon, and cervical cancer.

Main Results:

• DenseNet121 achieved the highest validation accuracy at 99.94% with a loss of 0.0017.
• DenseNet121 demonstrated the lowest Root Mean Square Error (RMSE) for both training (0.036056) and validation (0.045826).
• The study successfully identified DenseNet121 as the top-performing model.

Conclusions:

• AI-based techniques, particularly deep learning, significantly enhance cancer detection accuracy.
• DenseNet121 shows exceptional capability for automated detection across multiple cancer types.
• This research highlights the potential of AI in improving early cancer diagnosis and patient outcomes.