U-Net: Convolutional Networks for Biomedical Image Segmentation (original) (raw)
Related papers
Biomedical Image Segmentation with Modified U-Net
Traitement du Signal
Image segmentation is an important field in image processing and computer vision, particularly in the development of methods to assist experts in the biomedical and medical fields. It plays a vital role in saving time and costs. One of the most successful and significant methods in image segmentation using deep learning is the U-Net model. In this paper, we propose U-Net11, a novel variant of U-Net that uses 11 convolutional layers and introduces some modifications to improve the segmentation performance. The classical U-Net model was developed and tested on three different datasets, outperforming the traditional U-Net approach. The U-Net11 model was evaluated for breast cancer segmentation, lung segmentation from CT images, and the nuclei segmentation dataset from the Data Science Bowl 2018 competition. These datasets are valuable due to their varying image quantities and the varying difficulty levels in segmentation tasks. The modified U-Net model has achieved Dice Similarity Coefficient scores of 69.09% on the breast cancer dataset, 95.02% on the lung segmentation dataset and 81.10% on the nuclei segmentation dataset, exceeding the performance of the classical U-Net model by 5%, 2% and 4% respectively. This difference in success rates is particularly significant for critical segmentation datasets. deep learning,
Diagnostics, 2020
During image segmentation tasks in computer vision, achieving high accuracy performance while requiring fewer computations and faster inference is a big challenge. This is especially important in medical imaging tasks but one metric is usually compromised for the other. To address this problem, this paper presents an extremely fast, small and computationally effective deep neural network called Stripped-Down UNet (SD-UNet), designed for the segmentation of biomedical data on devices with limited computational resources. By making use of depthwise separable convolutions in the entire network, we design a lightweight deep convolutional neural network architecture inspired by the widely adapted U-Net model. In order to recover the expected performance degradation in the process, we introduce a weight standardization algorithm with the group normalization method. We demonstrate that SD-UNet has three major advantages including: (i) smaller model size (23x smaller than U-Net); (ii) 8x fewer parameters; and (iii) faster inference time with a computational complexity lower than 8M floating point operations (FLOPs). Experiments on the benchmark dataset of the Internatioanl Symposium on Biomedical Imaging (ISBI) challenge for segmentation of neuronal structures in electron microscopic (EM) stacks and the Medical Segmentation Decathlon (MSD) challenge brain tumor segmentation (BRATs) dataset show that the proposed model achieves comparable and sometimes better results compared to the current state-of-the-art.