Deep learning, reusable and problem-based architectures for detection of consolidation on chest X-ray images

•ChestNet is proposed which is proportional to the size of the dataset for detecting consolidation in chest X-ray images.•ChestNet has two times fewer max-pooling layers than the VGG16 and DenseNet121 to preserves the features of the images.•An efficient pre-processing process is proposed to remove confounding variables and histogram difference between images.•An extra validation with a totally different dataset is performed to indicate the generality of the proposed model.