The paper introduces CE-Net, a context encoder network designed for 2D medical image segmentation. It addresses the limitations of U-Net, particularly the loss of spatial information due to consecutive pooling and strided convolutional operations. CE-Net consists of three main components: a feature encoder module, a context extractor module, and a feature decoder module. The feature encoder uses a pretrained ResNet block, while the context extractor includes a dense atrous convolution (DAC) block and a residual multi-kernel pooling (RMP) block. The DAC block captures high-level semantic features using multi-scale atrous convolutions, and the RMP block encodes multi-scale context information. The feature decoder module enhances the extracted features to produce the final segmentation mask. Experimental results on various tasks, including optic disc segmentation, retinal vessel detection, lung segmentation, cell contour segmentation, and retinal OCT layer segmentation, demonstrate that CE-Net outperforms state-of-the-art methods. The paper also includes ablation studies to validate the effectiveness of each component in CE-Net.The paper introduces CE-Net, a context encoder network designed for 2D medical image segmentation. It addresses the limitations of U-Net, particularly the loss of spatial information due to consecutive pooling and strided convolutional operations. CE-Net consists of three main components: a feature encoder module, a context extractor module, and a feature decoder module. The feature encoder uses a pretrained ResNet block, while the context extractor includes a dense atrous convolution (DAC) block and a residual multi-kernel pooling (RMP) block. The DAC block captures high-level semantic features using multi-scale atrous convolutions, and the RMP block encodes multi-scale context information. The feature decoder module enhances the extracted features to produce the final segmentation mask. Experimental results on various tasks, including optic disc segmentation, retinal vessel detection, lung segmentation, cell contour segmentation, and retinal OCT layer segmentation, demonstrate that CE-Net outperforms state-of-the-art methods. The paper also includes ablation studies to validate the effectiveness of each component in CE-Net.