Pancreatic cancer is globally recognized as the most severe forms, because of notably low 5-year of survival rate. The majority cases are attributed to this aggressive cancer, underscoring the critical need for early detection. AI, along with its subfields like machine learning and deep learning, has immense potential to bring about positive changes in human lives. Through various medical imaging techniques, we can visualize the internal structure of the human body. This article introduces an approach for pancreatic cancer detection in CT scan images through the integration of different techniques, with CNN, Naive Bayes, AdaBoost, and other image processing methods. A CNN is a form of artificial neural network primarily utilized for image recognition and processing because of its capability to identify patterns within images. Naive Bayes is a straightforward method for constructing classifiers, which are models that assign class labels to instances of a problem, with these class labels being selected from a finite set, along with other techniques for processing images. AdaBoost, abbreviated from Adaptive Boosting, is a supervised learning technique employed to classify data by amalgamating numerous weak or base learners (like decision trees) into a powerful learner. AdaBoost operates by adjusting the weights of instances in the training dataset based on the accuracy of prior classifications. Employing different activation functions, such as rectified linear unit and sigmoid, are used to analyze the image. Dropout function is used to counter the overfitting problem. Different urinal biomarkers are used in the diagnosis of pancreatic tumor. Classification is executed using CNN, Naive Bayes algorithms, and AdaBoost. The result is measured by calculating accuracy, precision, recall, and F1-score for the classification algorithms. CNN algorithm demonstrating superior accuracy. Despite the challenges in pancreatic cancer diagnosis, particularly due to its severity, this approach holds promise for improving early detection capabilities through machine learning and image processing.