• Corpus ID: 232428285

Classification of Hematoma: Joint Learning of Semantic Segmentation and Classification

@article{Hirano2021ClassificationOH,
  title={Classification of Hematoma: Joint Learning of Semantic Segmentation and Classification},
  author={Hokuto Hirano and Tsuyoshi Okita},
  journal={ArXiv},
  year={2021},
  volume={abs/2103.17172}
}
脳血腫は6-24時間に急成長を遂げ, 予測を誤り脳外 科医が手術しなければ命を落とす. しかし, 脳血腫には急成長 をする型と急成長に至らない型の2通りが存在し, CTなどによ る検査を用いて急成長にいたる脳血腫か否かを人工知能で分 析する技術の開発を行う. この脳血腫の分類問題においては, 正例の少ない分類問題であること, 血腫の形状が不定であるこ と、その他, 不均衡分類問題, 共変量シフト, 少量データ問題, 疑似相関問題などさまざまなものが存在する. このため, 単純 なVGGなどのCNNを用いた分類においては精度を出すことが 難しい. そこで, 本論文においては, 血腫のセマンティックセグ メンテーションと分類とのジョイント学習を行う方法を提案し て, 性能を評価した. 

Figures from this paper

References

SHOWING 1-10 OF 10 REFERENCES
U-Net: Convolutional Networks for Biomedical Image Segmentation
TLDR
It is shown that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks.
Fully Convolutional Networks for Semantic Segmentation
TLDR
It is shown that convolutional networks by themselves, trained end- to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation.
Contrast Extravasation on CT Predicts Mortality in Primary Intracerebral Hemorrhage
TLDR
Active contrast extravasation on CT in patients with primary ICH independently predicts mortality and hematoma growth.
Association Between Hypodensities Detected by Computed Tomography and Hematoma Expansion in Patients With Intracerebral Hemorrhage.
TLDR
Hypodensities within an acute ICH detected on an NCCT scan may predict hematoma expansion, independent of other clinical and imaging predictors, and serve as a useful addition to clinical algorithms for determining the risk of and treatment stratification for hematomas expansion.
Intracerebral Hematoma Morphologic Appearance on Noncontrast Computed Tomography Predicts Significant Hematoma Expansion
TLDR
Fluid levels, density heterogeneity, and margin irregularity on noncontrast CT are associated with hematoma expansion at 24 hours, and these markers may assist in prediction of outcomes in scenarios where CT angiography is not readily available and may be of future help in refining the predictive value of the CTAngiography spot sign.
Density and Shape as CT Predictors of Intracerebral Hemorrhage Growth
TLDR
Density heterogeneity independently predicted ICH growth using some definitions, and large ICHs were significantly more irregular in shape, heterogeneous in density, and had greater growth.
Hematoma enlargement in spontaneous intracerebral hemorrhage.
TLDR
It is concluded that patients admitted to a hospital within 6 hours of onset of ICH, especially those admitted within 2 hours, and patients with liver dysfunction or irregularly shaped large hematomas should be closely observed for at least 6 hours after onset in preparation for emergency surgery, since the risk of hematoma growth in these circumstances is high.
Black Hole Sign: Novel Imaging Marker That Predicts Hematoma Growth in Patients With Intracerebral Hemorrhage
TLDR
The CT black hole sign could be used as a simple and easy-to-use predictor for early hematoma growth in patients with intracerebral hemorrhage.
PRE- DICT/Sunnybrook ICH CTA Study Group. Intracerebral Hematoma Morphologic Appearance on Noncontrast Computed Tomography Predicts Significant Hematoma
  • Expansion. Stroke
  • 2015
Recombinant Activated Factor VII Intracerebral Hemorrhage Trial Investigators. Density and shape as CT predictors of intracerebral hemorrhage growth. Stroke
  • 2009