The methods using attention module have been studied recently on image processing using Convolution Neural Network (CNN). The main purpose of CNN, where the method is applied, is to emphasize important features and weaken less important ones. From this perspective, we propose Compression and Intensity modules in order to boost the representation of feature map, by focusing on pixel-wise spatial attention. For each pixel, the importance of the spatial information which the feature possesses is identified and enhanced, so that an efficient segmentation task can be performed. The performance of the proposed module with state-of-the-art CNN models outperformed other recent attention modules for the brain MRI segmentation evaluation on MRBrainS18.