This document summarizes a research paper that proposes a new model called mmsDCNN-DenseCRF for semantic segmentation of remote sensing imagery. The mmsDCNN-DenseCRF model combines a modified multiscale deformable convolutional neural network (mmsDCNN) with a dense conditional random field (DenseCRF). The mmsDCNN generates a preliminary segmentation map capturing multiscale features. A multi-level DenseCRF then optimizes the mmsDCNN output using superpixel-level and pixel-level context to produce the final segmentation result. Experiments on standard datasets demonstrate the model achieves state-of-the-art performance for remote sensing image semantic segmentation.