Automatic segmentation of kidney and liver tumors in CT images

Abstract

Automatic segmentation of hepatic lesions in computed tomography (CT) images is a challenging task to perform due to heterogeneous, diffusive shape of tumors and complex background. To address the problem more and more researchers rely on assistance of deep convolutional neural networks (CNN) with 2D or 3D type architecture that have proven to be effective in a wide range of computer vision tasks, including medical image processing. In this technical report, we carry out research focused on more careful approach to the process of learning rather than on complex architecture of the CNN. We have chosen MIC-CAI 2017 LiTS dataset for training process and the public 3DIRCADb dataset for validation of our method. The proposed algorithm reached DICE score 78.8% on the 3DIRCADb dataset. The described method was then applied to the 2019 Kidney Tumor Segmentation (KiTS-2019) challenge, where our single submission achieved 96.38% for kidney and 67.38% for tumor Dice scores.

Kidney Tumor Semantic Segmentation Using Deep Learning: A Survey of State-of-the-Art

“…These patches are fed into a 3D CNN that is profoundly linked, and additional post-processing is conducted using conditional random fields. Efremova et al [ 45 ] and Shen et al [ 135 ] applied U-Net and 3D U-Net to the job of kidney tumor segmentation, with all techniques achieving much better results than conventional methods [ 51 ]. The next step was to analyze images from two distinct viewpoints and then link them via two connected networks.…”

“…One-stage methods [ 39 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 ] are designed to predict the multi-class results directly from whole images. Myronenko et al [ 44 ], from arterial phase abdominal 3D CT images, presented an end-to-end boundary aware fully CNN for accurate kidney and kidney tumor semantic segmentation.…”

“…Myronenko et al [ 44 ], from arterial phase abdominal 3D CT images, presented an end-to-end boundary aware fully CNN for accurate kidney and kidney tumor semantic segmentation. Efremova et al [ 45 ] combined U-Net and LinkNet-34 with ImageNet-trained ResNet-34 to decrease the convergence time and overfitting. Their model has shown success in a wide range of computer vision applications, including medical image processing.…”

EfficientNet family U-Net models for deep learning semantic segmentation of kidney tumors on CT images

“…Authors use a fully convolutional neural network to semantically segment kidney and tumor areas using KiTS19 challenge dataset contrast-enhanced CT images. The authors Geethanjali and Dinesh (2021) a novel Attention U-Net model used an attention mechanism and modified U-Net architecture to segment kidney tumors from CT data with 0.86 accuracies (Efremova et al, 2019). The research proposes an automated segmentation method for locating kidney and liver cancers in CT scans using a fully convolutional neural network architecture.…”