Cascaded deep convolutional encoder-decoder neural networks for efficient liver tumor segmentation

Budak, Ümit; Guo, Yanhui; Tanyıldızı, Erkan; Şengür, Abdulkadir

doi:10.1016/j.mehy.2019.109431

Cited by 104 publications

(56 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…And those images were preprocessed using the same preprocessing technique that was implemented on the training data. The result of the network was evaluated using the respective ground truths of the images and the comparison result of this algorithm with works of Christ et al who had used a cascaded deep neural network with a 3D conditional random fields to segment liver and its lesions [12], Liu et al who came up with GIU-Net that combines the improved UNet with the graph cut algorithm for segmenting liver sequence images [10] and lastly with Budak et al who developed two cascaded encoder-decoder convolutional neural network for the segmentation of liver and its tumor [15], were also included.…”

Section: A Test Results For Liver Segmentationmentioning

confidence: 99%

“…The output of the first network becomes the input for the next network. They achieved the average DSC value of 0.9522 and 0.634 on liver and tumor segmentation respectively [15].…”

Section: Related Work In the Area Of The Researchmentioning

confidence: 98%

“…Images that were used to train and test liver and liver cancer segmentation algorithm developed by this thesis were taken from two publicly available datasets, 3Dircadb01 (3D Image Reconstruction for Comparison of Algorithm Database) [17] and LITS (Liver Tumor Segmentation) Challenge [18]. The 3DIRCADb dataset is challenging to utilize since there is a high variety of data and the liver and tumor complexity [15]. Table 1 shows detail information about the two datasets.…”

Section: A Datasetsmentioning

confidence: 99%

“…Researchers had used CNN and its extensions, fully connected layer and UNet, for liver and tumor segmentation [10]- [15]. As our knowledge up to now, maximum dice score obtained for liver and tumor segmentation are 0.9522 and 0.63 respectively.…”

Section: Introductionmentioning

confidence: 98%

See 3 more Smart Citations

Deep Learning Based Liver Cancer Segmentation from Computed Tomography Images

Ayalew

Fante

Aliy

2020

Preprint

View full text Add to dashboard Cite

Background: Liver cancer is the sixth most common cancer worldwide. According to WHO data in 2017, the liver cancer death in Ethiopia reached 1040 (0.16%) from all cancer deaths. Hepatocellular carcinoma (HCC), primary liver cancer causes the death of around 700,000 people each year worldwide and this makes it the third leading cause of cancer death. HCC is occurred due to cirrhosis and hepatitis B or C viruses. Liver cancer mostly diagnosed with a computed tomography (CT) scan. But, the detection of the tumor from the CT scan image is difficult since tumors have similar intensity with nearby tissues and may have a different appearance depending on their type, state, and equipment setting. Nowadays deep learning methods have been used for the segmentation of liver and its tumor from the CT scan images and they are more efficient than those traditional methods. But, they are computationally expensive and need many labeled samples for training, which are difficult in the case of biomedical images. Results: A deep learning-based segmentation algorithm is employed for liver and tumor segmentation from abdominal CT scan images. Three separate UNet models, one for liver segmentation and the others two for tumor segmentation from the segmented liver and directly from the abdominal CT scan image were used. A dice score of 0.96 was obtained for liver segmentation. And a dice score of 0.74 and 0.63 was obtained for segmentation of tumor from the liver and from abdominal CT scan image respectively. Conclusion: The research improves the liver tumor segmentation that will help the physicians in the diagnosis and detection of liver tumors and in designing a treatment plan for the patient. And for the patient, it increases the patients’ chance of getting treatment and decrease the mortality rate due to liver cancer.

show abstract

Section: A Test Results For Liver Segmentationmentioning

confidence: 99%

“…The output of the first network becomes the input for the next network. They achieved the average DSC value of 0.9522 and 0.634 on liver and tumor segmentation respectively [15].…”

Section: Related Work In the Area Of The Researchmentioning

confidence: 98%

Section: A Datasetsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 98%

See 2 more Smart Citations

Deep Learning Based Liver Cancer Segmentation from Computed Tomography Images

Ayalew

Fante

Aliy

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Besides having a healthy diet and being physically active, an early detection and intervention is also critical in mitigating the risk of liver cancer. Currently, with the rapid development of medical imaging technology, CT and MRI medical imaging examinations have been widely used in clinical applications to monitor the liver structure and state for diagnosis and treatment of liver cancer [4]. However, manually analyzing detected imaging slices is really a time-consuming and error-prone task to conduct for physicians and radiologists alike and there often exist some inter-observer variations for this kind of pixel-level labelling tasks [5].…”

Section: Introductionmentioning

confidence: 99%

Correlation Matters: Multi-scale Fine-Grained Contextual Information Extraction for Hepatic Tumor Segmentation

Pang

et al. 2020

Advances in Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

Automatic tumor segmentation has been used as a diagnostic aid in the identification of diseases such as tumors from liver CT scans, and their treatment. Owing to their success in computer vision tasks, the state-of-the-art Fully Convolutional Networks (FCNs) or U-Net based models have often been employed in many recent studies for automatic tumor segmentation to learn numerous weight-shared convolutional kernels and extract various semantic features. However, the correlation between different tumor regions in feature maps cannot be easily captured due to the lack of contextual dependencies, which in turn limits the representative capability of the adopted models and thus affects the accuracy of tumor segmentation results. To resolve this issue, we propose a novel framework for segmentation of tumors in liver CT scans, which can explicitly extract multi-scale fine-grained contextual information by adaptively aggregating local features with their global dependencies. The proposed multi-scale framework features a light model with a very few additional parameters, and also its visualization capability significantly boosts networks' interpretability. Experimental results on a real-world liver tumor CT dataset illustrate that the proposed framework achieves the state-of-the-art performance in terms of a number of widely used evaluation criteria for the hepatic tumor segmentation task.

show abstract

Optimized segmentation and classification for liver tumor segmentation and classification using opposition‐based spotted hyena optimization

Rela

Rao

Reddy

2020

Int J Imaging Syst Tech

View full text Add to dashboard Cite

In today's world, liver cancers are one of the mainly popular cancers occurring in the human body. The greater part of liver carcinomas is more prone to alcohol‐related hepatitis and cirrhosis conditions. Moreover, there is another form of cancer namely, metastatic liver cancer, where the tumor is initiated from other organs and extends to the liver. Early and premature diagnosis of liver cancer is necessary as it tends to improvise life expectancy. Nowadays, discriminating the liver and tumor parts from medical images with the aid of completely automated computer‐aided software is a more challenging task, since the liver disease can vary from person to person. This article attempts to implement the novel liver tumor segmentation and classification model using the optimization driven segmentation and classification model. The developed model carries out the task in five steps (a) Pre‐processing, (b) liver segmentation, (c) tumor segmentation, (d) feature extraction, and (e) classification. At first, the gathered CT images are subjected to pre‐processing with three steps that follow contrast enhancement by histogram equalization and noise filtering by the median filter. Next to the pre‐processing of the image, the liver is segmented from the CT abdominal image using adaptive thresholding pursued by level set segmentation. Further, a modified algorithm termed as Fuzzy Centroid‐based Region Growing Algorithm with tolerance optimization is developed and used for the tumor segmentation. From the segmented tumor image, three sets of features like gray‐level co‐occurrence matrix (GLCM), shape features, and local binary pattern (LBP) is utilized for the classifier training. In the classification side, two deep learning algorithms are used: recurrent neural network (RNN), and convolutional neural network (CNN). The tumor segmented image is given as input to the CNN, and the extracted features are given as input to the RNN. As an improvement, an optimized hybrid classifier is adopted for the hidden neuron optimization. Moreover, an improved meta‐heuristic algorithm called opposition‐based spotted hyena optimization (O‐SHO) is introduced to perform the optimized segmentation and classification. The experimental results show that the overall accuracy attained by the proposed model is efficient, less sensitive to noise, and performs superior on a diverse set of CT images.

show abstract

Cascaded deep convolutional encoder-decoder neural networks for efficient liver tumor segmentation

Cited by 104 publications

References 14 publications

Deep Learning Based Liver Cancer Segmentation from Computed Tomography Images

Deep Learning Based Liver Cancer Segmentation from Computed Tomography Images

Correlation Matters: Multi-scale Fine-Grained Contextual Information Extraction for Hepatic Tumor Segmentation

Optimized segmentation and classification for liver tumor segmentation and classification using opposition‐based spotted hyena optimization

Contact Info

Product

Resources

About