Weakly supervised deep learning-based intracranial hemorrhage localization

Nemcek, Jakub; Vičar, Tomáš; Jakubíček, Roman

doi:10.48550/arxiv.2105.00781

Cited by 4 publications

(9 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As part of our future work we plan to leverage the f-AnoGAN [30] network architecture for faster inference of color flow images. However, in future, if a few labeled images of hemorrhage are available, we may adopt a semisupervised learning approach where residual attention modules or attention maps may be used for detection of hemorrhage similar to [17], [18].…”

Section: Discussionmentioning

confidence: 99%

“…Recent advances in machine and deep learning methods for detection and localization tasks have shown promising results in a variety of healthcare applications including hemorrhage detection. However, most of existing hemorrhage detection approaches are either supervised [14]- [16], semi-supervised [17] or weakly supervised [18]. Also, most of the existing works are focused on brain intracranial hemorrhage (ICH) detection in head CT scans, and NCTH detection still remains largely unexplored.…”

Section: Introductionmentioning

confidence: 99%

“…By applying attention mechanisms, Wang et al proposed a semisupervised multi-task attention-based U-Net for ICH segmentation with small amount of labeled data [17]. Nemcek et al applied weakly supervised learning approach with only class label for ICH localization by detecting local extrema in attention maps from the CNN classification network [18]. However, obtaining hemorrhage data from a bleeding subject has its own challenges as timely intervention becomes a top priority in such cases.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automatic Hemorrhage Detection From Color Doppler Ultrasound Using a Generative Adversarial Network (GAN)-Based Anomaly Detection Method

Mitra

Qiu

MacDonald

et al. 2022

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

Hemorrhage control has been identified as a priority focus area both for civilian and military populations in the United States because exsanguination is the most common cause of preventable death in hemorrhagic injury. Non-compressible torso hemorrhage (NCTH) has high mortality rate and there are currently no broadly available therapies for NCTH outside of a surgical room environment. Novel therapies, which include High Intensity Focused Ultrasound (HIFU) have emerged as promising methods for hemorrhage control as they can non-invasively cauterize bleeding tissue deep within the body without injuring uninvolved regions. A major challenge in the application of HIFU with color Doppler US guidance is the interpretation and optimization of the blood flow images in real-time to identify the hemorrhagic focus. Today, this task requires an expert sonographer, limiting the utility of this therapy in non-clinical environments. In this work, we investigated the feasibility of an automated hemorrhage detection method using a Generative Adversarial Network (GAN) for anomaly detection that learns a manifold of normal blood flow variability and subsequently identifies anomalous flow patterns that fall outside the learned manifold. As an initial feasibility study, we collected ultrasound color Doppler images of femoral arteries in an animal model of vascular injury (N=11 pigs). Velocity information of the blood flow were extracted from the color Doppler images that were used for training and testing the anomaly detection network. Normotensive images from 8 pigs were used for training, and testing was performed on normotensive, immediately after injury, 10 minutes post-injury and 30 minutes post-injury images from 3 other pigs. The residual images or the reconstructed error maps show promise in detecting hemorrhages with an AUC of 0.90, 0.87, 0.62 immediately, 10 minutes post-injury and 30 minutes post-injury respectively with an overall AUC of 0.83.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Hemorrhage Detection From Color Doppler Ultrasound Using a Generative Adversarial Network (GAN)-Based Anomaly Detection Method

Mitra

Qiu

MacDonald

et al. 2022

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

show abstract

“…Following the trend in explainable artificial intelligence (XAI), attention mechanisms have been employed to both boost the detection accuracy and visually illustrate classification results (Salehinejad et al, 2021). Furthermore, very limited attempts were also made to apply the attention/class activation in weakly supervised brain lesion and hemorrhage segmentation (Wu et al, 2019;Nemcek et al, 2021). Specifically, Wu et al (2019) used refined 3D Class-Activation Maps (CAMs) to segment stroke lesions from the Ischemic Stroke Lesion Segmentation (ISLES) dataset (multi-spectral MRI), and achieved a 0.3827 mean Dice score.…”

Section: Related Workmentioning

confidence: 99%

Weakly Supervised Intracranial Hemorrhage Segmentation Using Hierarchical Combination of Attention Maps from a Swin Transformer

Rasoulian

Salari

Xiao

2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Intracranial hemorrhage (ICH) is a life-threatening medical emergency caused by various factors. Timely and precise diagnosis of ICH is crucial for administering effective treatment and improving patient survival rates. While deep learning techniques have emerged as the leading approach for medical image analysis and processing, the most commonly employed supervised learning often requires large, high-quality annotated datasets that can be costly to obtain, particularly for pixel/voxel-wise image segmentation. To address this challenge and facilitate ICH treatment decisions, we proposed a novel weakly supervised ICH segmentation method that leverages a hierarchical combination of head-wise gradient-infused self-attention maps obtained from a Swin transformer. The transformer is trained using an ICH classification task with categorical labels. To build and validate the proposed technique, we used two publicly available clinical CT datasets, namely RSNA 2019 Brain CT hemorrhage and PhysioNet. Additionally, we conducted an exploratory study comparing two learning strategies -binary classification and full ICH subtyping -to assess their impact on self-attention and our weakly supervised ICH segmentation framework. The proposed algorithm was compared against the popular U-Net with full supervision, as well as a similar weakly supervised approach using Grad-CAM for ICH segmentation. With a mean Dice score of 0.47, our technique achieved similar ICH segmentation performance as the U-Net and outperformed the Grad-CAM based approach, demonstrating the excellent potential of the proposed framework in challenging medical image segmentation tasks.

show abstract

“…The algorithm [21] works with 2D axial slices of the CT scan and determines the positional coordinates of each bleeding location. The goal was to create a CNNbased detector using data with classification annotations (i.e., the positions of the ICHs in the slices were unknown).…”

Section: We Ak Ly-super Vised Le Ar Nin G-based M Ethodmentioning

confidence: 99%

Analysis of Circulatory System Pathologies in Head Ct Data – Hemorrhage Localization

Nemcek¹,

Jakubíček²

2021

Clin Tech

View full text Add to dashboard Cite

Acute ischemic stroke and intracranial hemorrhages (ICH) represent critical situations for the patient. Rapid accurate diagnosis and therapy are required to prevent serious lifelong consequences or death. In the case of suspected head circulatory pathology, computed tomography (CT) is often the first choice among imaging techniques because of its availability, speed and reliability. In order to refine and speed up the diagnostic process, advanced algorithms implemented in computer aided diagnosis systems are currently being developed. This paper presents approaches to an automatic ICH localization as a part of a research project aimed at the development of machine learning methods for the analysis of circulatory disorders in head CT scans. Three designed deep learning-based algorithms are described and compared for prediction of the exact position of ICH within a 3D CT scan, and in two cases also for classification into the sub-types. An objective evaluation of the methods is presented along with a discussion of the results. Further possibilities for circulatory diseases analysis in head CT scans using modern algorithms are also discussed.

show abstract

Weakly supervised deep learning-based intracranial hemorrhage localization

Cited by 4 publications

References 16 publications

Automatic Hemorrhage Detection From Color Doppler Ultrasound Using a Generative Adversarial Network (GAN)-Based Anomaly Detection Method

Automatic Hemorrhage Detection From Color Doppler Ultrasound Using a Generative Adversarial Network (GAN)-Based Anomaly Detection Method

Weakly Supervised Intracranial Hemorrhage Segmentation Using Hierarchical Combination of Attention Maps from a Swin Transformer

Analysis of Circulatory System Pathologies in Head Ct Data – Hemorrhage Localization

Contact Info

Product

Resources

About