Pneumonia is a lung ailment caused by bacterial or viral infections marked by swelling and trouble breathing deeply. Chest X-rays are a typical way to diagnose, which is so important. This study examines the implementation of deep learning approaches for pneumonia identification through chest X-ray images. The dataset has images of chest X-rays arranged based on whether they are pneumonia cases or not, and these categories are further split into train, test, and validation folders. Providing a dataset comprising 5,863 JPG format X-rays, which are 4,273 Pneumonia and 1,583 Normal, the dataset is a worthy resource for training and verifying deep learning methods in distinguishing pneumonia from a normal X-ray. The methodology that takes place in training the deep learning model is a progressive process, starting with preprocessing steps and executing morphological operations and histogram equalization. Transfer learning comes into play by using pretrained models that have been made from various existing architectures such as MobileNetV2, DenseNet201, VGG19, and Xception, which are used to extract features from the images, either alone or in a hybrid model that integrates features from multiple models. The models performed adequately, with MobileNetV2 achieving the maximum accuracy of 97.43%, followed closely by DenseNet201 at 96.57% and Xception at 95.87%. VGG19 recorded the lowest accuracy at 92.39%. Combining MobileNetV2 and DenseNet201 into a hybrid model notably improved accuracy to 99.10%. In the context of Industry 5.0, these models serve as collaborative instruments for radiologists, enhancing their ability to evaluate images from X-rays with more accuracy and efficiency, so augmenting the role of the human expert rather than substitutes it.