Helicobacter pylori bacteria are the common causal agent for antral gastritis. Chronic conditions of this type of gastritis may start gastric cancer. It seems to be one of the primary reasons for deaths due to gastric cancer throughout the world. To deal with this, early precursory detection and treatment is the only way. In this regard, artificial intelligence and its correlated fields like machine learning as well as deep learning are playing a great role. In particular, for image analysis one of the Deep Learning models called Convolutional neural network (CNN) has been profusely studied and used since last decade. This research work aims to detect the infection at the antral gastric mucosa related to H. pylori with the help of this revolutionary artificial neural network. Activation functions exhibit a significant role in deciding CNN performance and accuracy. It also demonstrates the performance and accuracy of the model by a comparative study of specific activation functions in the CNN model.