Today, the recognition of brain tumors is considered an essential topic in medicine. Brain neoplasm is an acute form of cancer caused by abnormal and uncontrollable cell division. Recent advancements in medical imaging-based deep learning have greatly assisted the healthcare industry in diagnosing a variety of diseases. To achieve image recognition and visual learning, a deep convolutional neural network (CNN) has been selected for implementing brain neoplasm recognition. CNN is a widely used and typical machine learning algorithm. This paper introduces a CNN-based approach along with data augmentation for categorizing magnetic resonance imaging (MRI) brain scan images into natural and unnatural categories. The proposed model achieves close-to-real-time recognition without sacrificing performance. Additionally, this paper describes the steps involved in setting hyperparameters, as well as the entire pipeline of the suggested pattern. After testing the system several times to find the optimal configuration that produces statistically more trustworthy results, each hyperparameter is chosen. Empirical results demonstrate that the proposed model achieves an accuracy of 99.55%. This model exhibits a low level of complexity and delivers more effective, accurate results when compared to other pre-trained models.