Improved Patch-Based Automated Liver Lesion Classification by Separate Analysis of the Interior and Boundary Regions

Diamant, Idit; Hoogi, Assaf; Beaulieu, Christopher F.; Safdari, Mustafa; Klang, Eyal; Amitai, Michal Marianne; Greenspan, Hayit; Rubin, Daniel L.

doi:10.1109/jbhi.2015.2478255

Cited by 45 publications

(32 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Accuracy value for proposed method in three dataset is 98%, 97.65%, and 98.1%. Comparing to other methods, our system specificity is 0.94% superior to Chen et al, 27 4.21% superior to Diamant et al, 28 6.48% superior to Abdar et al 26 (A) and 70% superior to Abdar et al 26 (B). Sensitivity value for proposed method in three dataset is 95.82%, 95%, and 95.34%.…”

Section: Performance Comparisonmentioning

confidence: 57%

“…Our proposed method is compared with other optimizations for finding best DBN parameters such as GA, PSO, ACO and FA. Proposed method sensitivity is 7.86% superior to Chen et al, 27 7.8% superior to Diamant et al, 28 3.16% superior to Abdar et al 26 (A) and 21.59% superior to Abdar et al 26 (B). Our method is 4.71% higher than GA, 5.82% higher than PSO, 6.87% higher ACO, and 7.68% higher than FA.…”

Section: Performance Comparisonmentioning

confidence: 78%

“…The connections between all lower layers are directed, with the arrows pointed toward the layer that is closest to the data. Our method is compared with the existing techniques such as, Chen et al, 27 Diamant et al, 28 and Abdar et al 26 In Chen et al 27 Three-way decision support is utilized to the diagnosis of focal liver lesion. The above Figure 4 illustrates the Comparison of accuracy, sensitivity and specificity values that were obtained for proposal and other methods.…”

Section: Performance Comparisonmentioning

confidence: 99%

“…Two layers are connected by a matrix of symmetrical weights W. Thus, the final architecture of DBN has been developed using the parameters such as nodes, layers (visible layer, hidden layer and output layer) as well as weights and biases for the layers. New BoVW approach utilizes SVM for liver lesion classification in Diamant et al 28 Boosted C5.0 (A) and CHAID algorithms (B) are used in Abdar et al 26 (A and B) for liver disease detection. This experiment conducted with three datasets (ILPD, OsiriX, TCGA-LIHC) was fed to a GOA method to search for the best DBN parameters.…”

Section: Performance Comparisonmentioning

confidence: 99%

“…Threeway decision support method could balance well the risk and efficiency of decision making. Diamant et al 28 introduced an automatic liver lesion classification by separate analysis of the interior and boundary regions. By using an image patch representation of the interior and boundary regions of the lesions their approach improved over single-dictionary bag-of-visualwords (BoVW) methods.…”

Section: Related Workmentioning

confidence: 99%

See 4 more Smart Citations

Optimizing deep belief network parameters using grasshopper algorithm for liver disease classification

Renukadevi¹,

Karunakaran²

2019

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Image processing plays a vital role in many areas such as healthcare, military, scientific and business due to its wide variety of advantages and applications. Detection of computed tomography (CT) liver disease is one of the difficult tasks in the medical field. Hand crafted features and classifications are the two types of methods used in the previous approaches, to classify liver disease. But these classification results are not optimal. In this article, we propose a novel method utilizing deep belief network (DBN) with grasshopper optimization algorithm (GOA) for liver disease classification. Initially, the image quality is enhanced by preprocessing techniques and then features like texture, color and shape are extracted. The extracted features are reduced by utilizing the dimensionality reduction method like principal component analysis (PCA). Here, the DBN parameters are optimized using GOA for recognizing liver disease. The experiments are performed on the real time and open source CT image datasets which embraces normal, cyst, hepatoma, and cavernous hemangiomas, fatty liver, metastasis, cirrhosis, and tumor samples. The proposed method yields 98% accuracy, 95.82% sensitivity, 97.52% specificity, 98.53% precision, and 96.8% F-1 score in simulation process when compared with other existing techniques. K E Y W O R D S deep belief network (DBN), grasshopper optimization algorithm (GOA), liver disease classification, principal component analysis (PCA)

show abstract

Section: Performance Comparisonmentioning

confidence: 57%

Section: Performance Comparisonmentioning

confidence: 78%

Section: Performance Comparisonmentioning

confidence: 99%

Section: Performance Comparisonmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Optimizing deep belief network parameters using grasshopper algorithm for liver disease classification

Renukadevi¹,

Karunakaran²

2019

Int J Imaging Syst Tech

View full text Add to dashboard Cite

show abstract

Automatic liver tumor segmentation on multiphase computed tomography volume using SegNet deep neural network and K‐means clustering

Pattwakkar

Kamath

Nanjundappa

et al. 2022

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Liver and liver tumor segmentations are essential in computer-aided systems for diagnosing liver tumors. These systems must operate on multiphase computed tomography (CT) images instead of a single phase for accurate diagnosis for clinical applications. We have proposed a framework that can perform segmentation from quadriphasic CT data. The liver was segmented using a finetuned SegNet model and the liver tumor was segmented using the K-means clustering method coupled with a power-law transformation-based image enhancement technique. The best values for liver segmentation achieved were: Dice Coefficient = 96.46 ± 0.48%, Jaccard Index = 93.16 ± 0.89%, volumetric overlap error = 6.84 ± 0.89% and average symmetric surface distance = 0.59 ± 0.3 mm and the results for liver tumor delineation were Dice Coefficient = 85.07 ± 4.5%, Jaccard Index = 74.29 ± 6.8%, volumetric overlap error = 25.71 ± 6.8% and average symmetric surface distance = 1.14 ± 1.3 mm. The proposed liver segmentation method based on deep learning is fully automatic, robust, and effective for all phases. The image enhancement technique has shown promising results and aided in better liver tumor segmentation. The liver tumors were segmented satisfactorily; however, improvements concerning false positive reduction can further increase the accuracy.

show abstract

Classification of focal liver lesions in CT images using convolutional neural networks with lesion information augmented patches and synthetic data augmentation

Lee

Hong

et al. 2021

Medical Physics

View full text Add to dashboard Cite

Purpose We propose a deep learning method that classifies focal liver lesions (FLLs) into cysts, hemangiomas, and metastases from portal phase abdominal CT images. We propose a synthetic data augmentation process to alleviate the class imbalance and the Lesion INformation Augmented (LINA) patch to improve the learning efficiency. Methods A dataset of 502 portal phase CT scans of 1,290 FLLs was used. First, to alleviate the class imbalance and to diversify the training data patterns, we suggest synthetic training data augmentation using DCGAN‐based lesion mask synthesis and pix2pix‐based mask‐to‐image translation. Second, to improve the learning efficiency of convolutional neural networks (CNNs) for the small lesions, we propose a novel type of input patch termed the LINA patch to emphasize the lesion texture information while also maintaining the lesion boundary information in the patches. Third, we construct a multi‐scale CNN through a model ensemble of ResNet‐18 CNNs trained on LINA patches of various mini‐patch sizes. Results The experiments demonstrate that (a) synthetic data augmentation method shows characteristics different but complementary to those in conventional real data augmentation in augmenting data distributions, (b) the proposed LINA patches improve classification performance compared to those by existing types of CNN input patches due to the enhanced texture and boundary information in the small lesions, and (c) through an ensemble of LINA patch‐trained CNNs with different mini‐patch sizes, the multi‐scale CNN further improves overall classification performance. As a result, the proposed method achieved an accuracy of 87.30%, showing improvements of 10.81%p and 15.0%p compared to the conventional image patch‐trained CNN and texture feature‐trained SVM, respectively. Conclusions The proposed synthetic data augmentation method shows promising results in improving the data diversity and class imbalance, and the proposed LINA patches enhance the learning efficiency compared to the existing input image patches.

show abstract

Improved Patch-Based Automated Liver Lesion Classification by Separate Analysis of the Interior and Boundary Regions

Cited by 45 publications

References 35 publications

Optimizing deep belief network parameters using grasshopper algorithm for liver disease classification

Optimizing deep belief network parameters using grasshopper algorithm for liver disease classification

Automatic liver tumor segmentation on multiphase computed tomography volume using SegNet deep neural network and K‐means clustering

Classification of focal liver lesions in CT images using convolutional neural networks with lesion information augmented patches and synthetic data augmentation

Contact Info

Product

Resources

About