U-Net is a state-of-the-art, wildly used deep convolutional neural network for image segmentation. It consists of encoder-decoder style architecture and skip-connections between them to localize high resolution features. In this project we examined a possible improvement to the architecture by adding unsupervised learning method (auto-encoder) in U-Net architecture variant (U-Net with ResNet-34 as encoder). We used the data and evaluation system from a Kaggle competition – “Understanding Clouds from Satellite Images” and showed improvement in performance of a “baseline” model by implementing our solution. Furthermore, we demonstrate the capability to localize high resolution features for segmentation using data contributed by the auto-encoder, rather than the skip-connections.