JPWO2021070108A5 - - Google Patents

Download PDF

Info

Publication number
JPWO2021070108A5
JPWO2021070108A5 JP2022520838A JP2022520838A JPWO2021070108A5 JP WO2021070108 A5 JPWO2021070108 A5 JP WO2021070108A5 JP 2022520838 A JP2022520838 A JP 2022520838A JP 2022520838 A JP2022520838 A JP 2022520838A JP WO2021070108 A5 JPWO2021070108 A5 JP WO2021070108A5
Authority
JP
Japan
Prior art keywords
disease
detection model
layer
dimensional slices
convolutional
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP2022520838A
Other languages
Japanese (ja)
Other versions
JP2022553906A (en
Publication date
Priority claimed from US16/600,009 external-priority patent/US11195273B2/en
Application filed filed Critical
Publication of JP2022553906A publication Critical patent/JP2022553906A/en
Publication of JPWO2021070108A5 publication Critical patent/JPWO2021070108A5/ja
Pending legal-status Critical Current

Links

Claims (14)

疾患検出モデルを開発するための方法であって、
１つまたは複数の電子プロセッサを用いて、画像学習と、前記画像学習について生成された読影レポートからマイニングされた関連付けられた疾患ラベルとを使用して前記疾患検出モデルを訓練することであって、前記画像学習は３次元画像ボリュームの複数の２次元スライスのシーケンスを含み、前記疾患検出モデルは畳み込みニューラル・ネットワーク層および畳み込み長短期記憶層を含み、前記疾患検出モデルを訓練することは、
前記畳み込みニューラル・ネットワーク層を使用して前記複数の２次元スライスのそれぞれから特徴のセットを個別に抽出することと、
前記畳み込み長短期記憶層を使用して前記複数の２次元スライスのそれぞれについて前記畳み込みニューラル・ネットワーク層によって抽出された前記特徴のセットを順次処理することと、
疾患の確率を生成するために前記複数の２次元スライスのうちの順番が最後のものに関する前記畳み込み長短期記憶層からの出力を処理することと、
前記確率を前記疾患ラベルと比較することと、
前記比較に基づいて前記疾患検出モデルを更新することと、
を含む、前記訓練することと、
前記疾患検出モデルが訓練された後に、新たな画像学習の前記疾患の確率を生成するために前記新たな画像学習に前記疾患検出モデルを適用することと、
を含む、方法。 A method for developing a disease detection model, the method comprising:
training the disease detection model using one or more electronic processors using image learning and associated disease labels mined from interpretation reports generated for the image learning; the image training includes a sequence of a plurality of two-dimensional slices of a three-dimensional image volume, the disease detection model includes a convolutional neural network layer and a convolutional long short-term memory layer, and training the disease detection model comprises:
separately extracting a set of features from each of the plurality of two-dimensional slices using the convolutional neural network layer;
sequentially processing the set of features extracted by the convolutional neural network layer for each of the plurality of two-dimensional slices using the convolutional long short-term memory layer;
processing the output from the convolutional long short-term memory layer for the last in order of the plurality of two-dimensional slices to generate a probability of disease;
comparing the probability to the disease label;
updating the disease detection model based on the comparison;
said training, including;
After the disease detection model is trained, applying the disease detection model to the new image learning to generate a new image learning probability of the disease;
including methods. 前記疾患ラベルは、前記画像学習に関連付けられた患者が前記疾患と診断されたか否かのバイナリ・ラベルを含む、請求項１に記載の方法。 2. The method of claim 1, wherein the disease label includes a binary label of whether a patient associated with the image learning has been diagnosed with the disease. 前記疾患の前記確率は肺気腫の確率である、請求項１または２に記載の方法。 3. The method of claim 1 or 2, wherein the probability of the disease is the probability of emphysema. 前記複数の２次元スライスは、胸部を含む一連のコンピュータ断層撮影（ＣＴ）画像を含む、請求項１から３のいずれか一項に記載の方法。 4. A method according to any one of claims 1 to 3, wherein the plurality of two-dimensional slices comprises a series of computed tomography (CT) images comprising a thorax. 前記３次元画像ボリュームはアノテーションを含まない、請求項１から４のいずれか一項に記載の方法。 5. A method according to any one of claims 1 to 4, wherein the three-dimensional image volume does not include annotations. 前記疾患検出モデルは２つの双方向のユニットのペアを含み、各ユニットは前記畳み込み層および前記畳み込み長短期記憶層を含む、請求項１から５のいずれか一項に記載の方法。 6. A method according to any one of claims 1 to 5, wherein the disease detection model comprises a pair of two bidirectional units, each unit comprising the convolutional layer and the convolutional long short-term memory layer. 前記疾患の確率を生成するために前記複数の２次元スライスのうちの前記順番が最後のものに関する前記畳み込み長短期記憶層からの前記出力を処理することは、前記複数の２次元スライスのうちの前記順番が最後のものからシグモイド活性化型Ｄｅｎｓｅ層に単一の特徴のセットを出力することを含む、請求項１から６のいずれか一項に記載の方法。 Processing the output from the convolutional long short-term memory layer for the last in order of the plurality of two-dimensional slices to generate the probability of disease comprises: 7. A method according to any one of claims 1 to 6, comprising outputting a single set of features to the sigmoid-activated Dense layer from the last in the order. 疾患検出モデルを開発するためのシステムであって、
１つまたは複数の電子プロセッサを備え、前記１つまたは複数の電子プロセッサは、
請求項１から７のいずれか一項に記載の方法を実行するように構成されている、システム。 A system for developing a disease detection model, the system comprising:
one or more electronic processors, the one or more electronic processors comprising:
A system configured to carry out a method according to any one of claims 1 to 7. 前記畳み込み長短期記憶層を使用して前記複数の２次元スライスのそれぞれについて前記畳み込みニューラル・ネットワーク層によって抽出された前記特徴のセットを順次処理することは、前記３次元画像ボリュームにわたる空間パターンおよび変動を検出することを含む、請求項８に記載のシステム。 Sequentially processing the set of features extracted by the convolutional neural network layer for each of the plurality of two-dimensional slices using the convolutional long short-term memory layer may be configured to sequentially process the set of features extracted by the convolutional neural network layer for each of the plurality of two-dimensional slices. 9. The system of claim 8, comprising detecting. ２つの双方向のユニットのペアは、前記複数の２次元スライスの昇順で前記３次元画像ボリュームに３２個のフィルタを適用する第１のユニットと、前記複数の２次元スライスの降順で前記３次元画像ボリュームに３２個のフィルタを適用する第２のユニットと、前記昇順で前記３次元画像ボリュームに６４個のフィルタを適用する第３のユニットと、前記降順で前記３次元画像ボリュームに６４個のフィルタを適用する第４のユニットと、を含む、請求項８または９に記載のシステム。 A pair of two bidirectional units includes a first unit that applies 32 filters to the three-dimensional image volume in ascending order of the plurality of two-dimensional slices, and a first unit that applies thirty-two filters to the three-dimensional image volume in ascending order of the plurality of two-dimensional slices; a second unit applying 32 filters to the image volume; a third unit applying 64 filters to the 3D image volume in the ascending order; and a third unit applying 64 filters to the 3D image volume in the descending order. 10. The system according to claim 8 or 9, comprising a fourth unit for applying a filter. 前記１つまたは複数のプロセッサは、前記複数の２次元スライスのうちの前記順番が最後のものからシグモイド活性化型Ｄｅｎｓｅ層に単一の特徴のセットを出力することによって、前記疾患の確率を生成するために前記複数の２次元スライスのうちの前記順番が最後のものに関する前記畳み込み長短期記憶層からの前記出力を処理するように構成されている、請求項８から１０のいずれか一項に記載のシステム。 The one or more processors generate the disease probability by outputting a single set of features from the last in order of the plurality of two-dimensional slices to a sigmoid-activated dense layer. 11. According to any one of claims 8 to 10, the output from the convolutional long short-term storage layer for the last of the plurality of two-dimensional slices in order to The system described. 前記疾患検出モデルは、前記畳み込み長短期記憶層のための最大値プーリング層をさらに含む、請求項８から１１のいずれか一項に記載のシステム。 12. The system of any one of claims 8 to 11, wherein the disease detection model further comprises a maximum pooling layer for the convolutional long short-term memory layer. 請求項１から７のいずれか一項に記載の方法を、コンピュータに実行させる、コンピュータ・プログラム。 A computer program that causes a computer to execute the method according to any one of claims 1 to 7. 請求項１３に記載の前記コンピュータ・プログラムを、コンピュータ可読記憶媒体に記憶した、記憶媒体。 A storage medium storing the computer program according to claim 13 on a computer readable storage medium.
JP2022520838A 2019-10-11 2020-10-08 Systems, methods and programs for developing disease detection models Pending JP2022553906A (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US16/600,009 US11195273B2 (en) 2019-10-11 2019-10-11 Disease detection from weakly annotated volumetric medical images using convolutional long short-term memory
US16/600,009 2019-10-11
PCT/IB2020/059467 WO2021070108A1 (en) 2019-10-11 2020-10-08 Disease detection from weakly annotated volumetric medical images using convolutional long short-term memory

Publications (2)

Publication Number Publication Date
JP2022553906A JP2022553906A (en) 2022-12-27
JPWO2021070108A5 true JPWO2021070108A5 (en) 2023-10-03

Family

ID=75383038

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2022520838A Pending JP2022553906A (en) 2019-10-11 2020-10-08 Systems, methods and programs for developing disease detection models

Country Status (6)

Country Link
US (1) US11195273B2 (en)
JP (1) JP2022553906A (en)
CN (1) CN114503213B (en)
DE (1) DE112020004049T5 (en)
GB (1) GB2604503B (en)
WO (1) WO2021070108A1 (en)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11638569B2 (en) 2018-06-08 2023-05-02 Rutgers, The State University Of New Jersey Computer vision systems and methods for real-time needle detection, enhancement and localization in ultrasound
WO2020036968A1 (en) 2018-08-13 2020-02-20 Rutgers, The State University Of New Jersey Computer vision systems and methods for real-time localization of needles in ultrasound images
US11217350B2 (en) * 2019-06-27 2022-01-04 Retrace Labs Systems and method for artificial-intelligence-based dental image to text generation
US20210374384A1 (en) * 2020-06-02 2021-12-02 Nvidia Corporation Techniques to process layers of a three-dimensional image using one or more neural networks
EP3923183A1 (en) * 2020-06-11 2021-12-15 Tata Consultancy Services Limited Method and system for video analysis
CN113221277B (en) * 2021-05-14 2022-12-09 西安交通大学 Bearing performance degradation evaluation method and system based on digital twin model
CN113469229A (en) * 2021-06-18 2021-10-01 中山大学孙逸仙纪念医院 Method and device for automatically labeling breast cancer focus based on deep learning
CN113723210A (en) * 2021-08-05 2021-11-30 中冶南方城市建设工程技术有限公司 Method and system for detecting leakage edge equipment of water supply pipe in intelligent park
CN113966999A (en) * 2021-10-28 2022-01-25 中山大学 Sorafenib drug resistance prediction method and device and storage medium
US20230186463A1 (en) * 2021-12-09 2023-06-15 Merative Us L.P. Estimation of b-value in prostate magnetic resonance diffusion weighted images
US20240152747A1 (en) * 2022-11-08 2024-05-09 UnitX, Inc. Three-dimensional spatial-channel deep learning neural network

Family Cites Families (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015054597A2 (en) * 2013-10-12 2015-04-16 H. Lee Moffitt Cancer Center And Research Institute, Inc. Systems and methods for diagnosing tumors in a subject by performing a quantitative analysis of texture-based features of a tumor object in a radiological image
WO2017151757A1 (en) 2016-03-01 2017-09-08 The United States Of America, As Represented By The Secretary, Department Of Health And Human Services Recurrent neural feedback model for automated image annotation
US10354168B2 (en) * 2016-04-11 2019-07-16 A2Ia S.A.S. Systems and methods for recognizing characters in digitized documents
US9767557B1 (en) 2016-06-23 2017-09-19 Siemens Healthcare Gmbh Method and system for vascular disease detection using recurrent neural networks
CN106250707A (en) 2016-08-12 2016-12-21 王双坤 A kind of based on degree of depth learning algorithm process head construction as the method for data
US10417788B2 (en) * 2016-09-21 2019-09-17 Realize, Inc. Anomaly detection in volumetric medical images using sequential convolutional and recurrent neural networks
US10304198B2 (en) 2016-09-26 2019-05-28 Siemens Healthcare Gmbh Automatic medical image retrieval
CN107103187B (en) 2017-04-10 2020-12-29 四川省肿瘤医院 Lung nodule detection grading and management method and system based on deep learning
WO2018204311A1 (en) * 2017-05-01 2018-11-08 The Charles Stark Draper Laboratory, Inc. Deep learning architecture for cognitive examination subscore trajectory prediction in alzheimer's disease
CN107392895A (en) 2017-07-14 2017-11-24 深圳市唯特视科技有限公司 A kind of 3D blood vessel structure extracting methods based on convolution loop network
US10595727B2 (en) 2018-01-25 2020-03-24 Siemens Healthcare Gmbh Machine learning-based segmentation for cardiac medical imaging
CN108281183A (en) * 2018-01-30 2018-07-13 重庆大学 Cervical smear image diagnostic system based on convolutional neural networks and transfer learning
CN108470359A (en) 2018-02-11 2018-08-31 艾视医疗科技成都有限公司 A kind of diabetic retinal eye fundus image lesion detection method
CN108717869B (en) * 2018-05-03 2021-08-13 中国石油大学(华东) Auxiliary system for diagnosing diabetic retinal complications based on convolutional neural network
CN109102491B (en) 2018-06-28 2021-12-28 武汉楚精灵医疗科技有限公司 Gastroscope image automatic acquisition system and method
CN109147937A (en) 2018-07-31 2019-01-04 中国科学院深圳先进技术研究院 Rehabilitation prediction technique and Related product based on image
KR102135477B1 (en) 2018-08-31 2020-07-17 엔에이치엔 주식회사 Method and system for image automatic classification
CN109410194B (en) 2018-10-19 2022-03-29 山东省计算中心（国家超级计算济南中心） Esophageal cancer pathological image processing method based on deep learning
CN109585020A (en) * 2018-11-27 2019-04-05 华侨大学 A kind of model that disease risks are predicted with convolutional neural networks
CN109635246B (en) 2018-12-06 2021-07-13 西南交通大学 Multi-attribute data modeling method based on deep learning
CN110188936B (en) 2019-05-23 2021-11-05 浙江大学 Short-term traffic flow prediction method based on multi-factor space selection deep learning algorithm

Similar Documents

Publication Publication Date Title
GB2604503A (en) Disease detection from weakly annotated volumetric medical images using convolutional long short-term memory
US10417788B2 (en) Anomaly detection in volumetric medical images using sequential convolutional and recurrent neural networks
Jin et al. AI-assisted CT imaging analysis for COVID-19 screening: Building and deploying a medical AI system in four weeks
Zhang et al. Classification of lung nodules based on CT images using squeeze-and-excitation network and aggregated residual transformations
CN111429421B (en) Model generation method, medical image segmentation method, device, equipment and medium
CN108171232A (en) The sorting technique of bacillary and viral children Streptococcus based on deep learning algorithm
CN109614874B (en) Human behavior recognition method and system based on attention perception and tree skeleton point structure
Khoiriyah et al. Convolutional neural network for automatic pneumonia detection in chest radiography
JPWO2021070108A5 (en)
CN113592809A (en) Pneumonia image detection system and method based on channel attention residual error network
Xue et al. Covid-19 diagnostic using 3d deep transfer learning for classification of volumetric computerised tomography chest scans
Zhang et al. Pulmonary nodule detection based on 3D feature pyramid network with incorporated squeeze‐and‐excitation‐attention mechanism
Aria et al. ADA-COVID: adversarial deep domain adaptation-based diagnosis of COVID-19 from lung CT scans using triplet embeddings
Zunair et al. Estimating Severity from CT Scans of Tuberculosis Patients using 3D Convolutional Nets and Slice Selection.
Zhu et al. Mvc-net: Multi-view chest radiograph classification network with deep fusion
US20220415021A1 (en) Data augmentation for a machine learning method
Muljo et al. Handling severe data imbalance in chest X-Ray image classification with transfer learning using SwAV self-supervised pre-training
Saikrishna et al. Pneumonia Detection Using Deep Learning Algorithms
CN116778527A (en) Human body model construction method, device, equipment and storage medium
Singh et al. Automated detection of lung cancer using transfer learning based deep learning
Esmaeili et al. Diagnosis of Covid-19 disease by combining hand-crafted and deep-learning methods on ultrasound data
Monshi et al. Distributed Deep Learning for Multi-Label Chest Radiography Classification.
CN113487575B (en) Method, apparatus, device and readable storage medium for training medical image detection model
CN117393100B (en) Diagnostic report generation method, model training method, system, equipment and medium
Keerthana Detection and Classification of Type 2 Diabetes using Machine Learning Techniques