Basic Ensembles of Vanilla-Style Deep Learning Models Improve Liver Segmentation From CT Images

Kavur, Ali Emre; Kuncheva, Ludmila I.; Selver, M. Alper

doi:10.1201/9781003215141-3

Cited by 14 publications

(3 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Drawing inspiration from the success of ensemble methods in classification tasks [58] and segmentation competitions [59], [60], we investigated their potential in context of MRI reconstruction. Employing three fundamental combination methods [61] (Min/Max Combiner, Averaging Combiner, and Weighted Averaging), we evaluated their efficacy in improving the reconstruction process. The results, as presented in Table 3, reveal that, surprisingly, basic ensembles yielded slightly higher performance in the reconstruction task.…”

Section: A Quantitative Reconstruction Resultsmentioning

confidence: 99%

Deep Learning-Based Meniscus Tear Detection From Accelerated MRI

Harman,

Selver,

Baris

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Magnetic Resonance Imaging (MRI) is an exceptional diagnostic tool known for its ability to provide superior soft tissue contrast. Despite its proven efficacy, traditional MRI methods have some inherent limitations, including relatively longer scan times and the requirement for specialized expertise in data analysis. These factors can, at times, create challenges in the widespread adoption of MRI in certain research and clinical scenarios. Deep learning (DL) methods for MRI reconstruction and analysis offer a promising solution to address this issue. While these DL techniques have been validated using standard image quality metrics, they fall short in assessing clinically relevant details. This hinders their clinical reliability and practical application. Although radiologists have previously compared the diagnostic equivalence of accelerated DL-reconstructed images to conventional ones for evaluating knee internal derangement, it remains uncertain whether DL detectors can accurately identify clinically important details when reconstruction models are used. Furthermore, the artifacts or hallucinations generated by DL reconstructions in knee MRI have not been examined in a clinical setting. Thus, this study aims to determine the performance of DL detectors on retrospectively accelerated DL-reconstructed knee MRI, comparing them to conventional imaging and expert evaluations for detecting meniscal tears. Our investigation yields three significant contributions. First, an in-depth analysis of DL reconstruction highlights the presence of hallucinations in the femur, tibia, and false positive artifacts, indicating that the overall reconstruction quality does not directly affect pathological features. Second, the results of DL detectors demonstrate that their performance aligns well with image quality assessment metrics and expert scores. This finding validates the reliability of the detection outcomes. Finally, we propose an integrated (i.e. reconstruction + detection) process for meniscal tears on fastMRI+ data and achieved state-of-the-art results with average precision scores of 0.69 and 0.67 at 4− and 8-fold accelerations, respectively.INDEX TERMS MRI reconstruction, meniscal tear detection, and deep learning.

show abstract

Section: A Quantitative Reconstruction Resultsmentioning

confidence: 99%

Deep Learning-Based Meniscus Tear Detection From Accelerated MRI

Harman,

Selver,

Baris

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…In recent times ensembling is a popular concept to improve the results for discriminative CNNs, particularly for image classification (Neena & Geetha, 2018), object detection (X. Wang & Gupta, 2015), or medical image segmentation tasks (Altaf et al, 2021; Kavur, Gezer, et al, 2020; Kavur, Kuncheva, et al, 2020; Menze et al, 2014). Ensembling of GANs has also been experimented with for imbalanced image classification (Ermaliuc et al, 2021; Huang et al, 2020).…”

Section: Methodsmentioning

confidence: 99%

Long future frame prediction using optical flow‐informed deep neural networks for enhancement of robotic teleoperation in high latency environments

Moniruzzaman

Rassau

Chai

et al. 2022

Journal of Field Robotics

View full text Add to dashboard Cite

High latency in teleoperation has a significant negative impact on operator performance. While deep learning has revolutionized many domains recently, it has not previously been applied to teleoperation enhancement. We propose a novel approach to predict video frames deep into the future using neural networks informed by synthetically generated optical flow information. This can be employed in teleoperated robotic systems that rely on video feeds for operator situational awareness. We have used the image-to-image translation technique as a basis for the prediction of future frames. The Pix2Pix conditional generative adversarial network (cGAN) has been selected as a base network. Optical flow components reflecting real-time control inputs are added to the standard RGB channels of the input image. We have experimented with three data sets of 20,000 input images each that were generated using our custom-designed teleoperation simulator with a 500-ms delay added between the input and target frames. Structural Similarity Index Measures (SSIMs) of 0.60 and Multi-SSIMs of 0.68 were achieved when training the cGAN with three-channel RGB image data. With the five-channel input data (incorporating optical flow) these values improved to 0.67 and 0.74, respectively. Applying Fleiss' κ gave a score of 0.40 for three-channel RGB data, and 0.55 for fivechannel optical flow-added data. We are confident the predicted synthetic frames are of sufficient quality and reliability to be presented to teleoperators as a video feed that will enhance teleoperation. To the best of our knowledge, we are the first to attempt to reduce the impacts of latency through future frame prediction using deep neural networks.

show abstract

“…To further improve the performance, (Kavur et al. 2022 ) proposed to combine four neural networks, U-Net, Deepmedic, V-Net, and Dense V-Networks. (Xie et al.…”

Section: Related Workmentioning

confidence: 99%

Sd-net: a semi-supervised double-cooperative network for liver segmentation from computed tomography (CT) images

Huang,

Luo,

et al. 2024

J Cancer Res Clin Oncol

View full text Add to dashboard Cite

Introduction The automatic segmentation of the liver is a crucial step in obtaining quantitative biomarkers for accurate clinical diagnosis and computer-aided decision support systems. This task is challenging due to the frequent presence of noise and sampling artifacts in computerized tomography (CT) images, as well as the complex background, variable shapes, and blurry boundaries of the liver. Standard segmentation of medical images based on full-supervised convolutional networks demands accurate dense annotations. Such a learning framework is built on laborious manual annotation with strict requirements for expertise, leading to insufficient high-quality labels. Methods To overcome such limitation and exploit massive weakly labeled data, we relaxed the rigid labeling requirement and developed a semi-supervised double-cooperative network (SD- Net). SD-Net is trained to segment the complete liver volume from preoperative abdominal CT images by using limited labeled datasets and large-scale unlabeled datasets. Specifically, to enrich the diversity of unsupervised information, we construct SD-Net consisting of two collaborative network models. Within the supervised training module, we introduce an adaptive mask refinement approach. First, each of the two network models predicts the labeled dataset, after which adaptive mask refinement of the difference predictions is implemented to obtain more accurate liver segmentation results. In the unsupervised training module, a dynamic pseudo-label generation strategy is proposed. First each of the two models predicts unlabeled data and the better prediction is considered as pseudo-labeling before training. Results and discussion Based on the experimental findings, the proposed method achieves a dice score exceeding 94%, indicating its high level of accuracy and its suitability for everyday clinical use.

show abstract

Basic Ensembles of Vanilla-Style Deep Learning Models Improve Liver Segmentation From CT Images

Cited by 14 publications

References 38 publications

Deep Learning-Based Meniscus Tear Detection From Accelerated MRI

Deep Learning-Based Meniscus Tear Detection From Accelerated MRI

Long future frame prediction using optical flow‐informed deep neural networks for enhancement of robotic teleoperation in high latency environments

Sd-net: a semi-supervised double-cooperative network for liver segmentation from computed tomography (CT) images

Contact Info

Product

Resources

About