EP4232998A1 - Procede de traitement d'image radiologique - Google Patents
Procede de traitement d'image radiologiqueInfo
- Publication number
- EP4232998A1 EP4232998A1 EP21794159.0A EP21794159A EP4232998A1 EP 4232998 A1 EP4232998 A1 EP 4232998A1 EP 21794159 A EP21794159 A EP 21794159A EP 4232998 A1 EP4232998 A1 EP 4232998A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- radiological
- anomaly
- influence
- image
- pixel
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10116—X-ray image
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Definitions
- the invention relates to a radiological image processing method, and falls within the field of medical devices for the exploitation of images obtained by X-ray imaging, and more specifically in the field of treatment of image and automatic learning to assist in the diagnosis of radiological examinations.
- a radiological anomaly corresponds to any visible sign in a radiological examination suggestive of a pathology in the patient.
- these anomalies can be opacities, condensations, calcifications, infiltrations...
- Recent work proposes providing diagnostic assistance, based for the most part on neural networks, more specifically on networks performing image classification.
- Such examples are described in "Abnormality Detection and Localization in Chest X-Rays Using Deep Convolutional Neural Networks", ArXiv:1705.09850 [Cs], 2017, http://arxiv.org/abs/1705.09850, by Islam, Mohammad Tariqul, Md Abdul Aowal, Ahmed Tahseen Minhaz, and Khalid Ashraf, and in the “CheXNet document: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning", ArXiv:1711.05225 [Cs, Stat], http://arxiv.org/abs/1711.05225, 2017, by Rajpurkar, Pranav, Jeremy Irvin, Kaylie Zhu, Brandon Yang, Hershel Mehta, Tony Duan, Daisy Ding, et al.
- An object of the invention is to overcome the problems mentioned above, and in particular to improve the post-processing of the digital images to better highlight the radiological anomalies detected, to facilitate the work of the radiologist and to reduce the diagnostic time. .
- a method for processing a radiological image /(x,y), in digital format comprising at least one radiological anomaly detected using a convolutional neural network having been trained to detect radiological anomalies on radiological examinations.
- the radiological image /(x,y) is characterized by the intensity of each of its pixels /(x,y), and by at least one radiological anomaly influence map C k (x,y) which attributes , for each pixel (x,y) of the radiological image /(x,y), a value representative of the proportion in which the pixel had an influence in the detection result of the radiological anomaly k.
- the method is implemented by computer and comprises the steps of:
- the method of the invention makes it possible to highlight the anomalies detected, to facilitate the work of the radiologist and to reduce the diagnostic time.
- each anomaly influence map C k (x, y) uses the following affine transformation: in which:
- C kn (x,y) is the normalized radiological anomaly influence map of the radiological anomaly influence map C k (x,y).
- the fusion C(x,y) uses an average of the normalized radiological anomaly influence maps C kn (x,y) weighted by the probability pk of presence of each anomaly in using the following relation: if
- represents the number of anomalies detected
- Pk represents the probability pk of presence of each anomaly, calculated by the convolutional neural network.
- Such a fusion has the advantage of having a single map to weight the contribution of each pixel of the image in the calculation of intensity histograms.
- the processing uses an intensity histogram equalization, the calculation of the histogram of which is modified by weighting the contribution of each pixel according to the following relationship:
- Such processing has the advantage of improving the contrast of the image, particularly that of the areas of the image presenting anomalies which have been detected, since the influence of the corresponding pixels in the calculation of the histogram is increased. .
- the processing can use a variation of the method proposed in the document "Automatic x-ray image contrast enhancement based on parameter auto-optimization”. Journal of Applied Clinical Medical Physics 18(6):218-23, by Qiu, Jianfeng, H. Harold Li, Tiezhi Zhang, Fangfang Ma, and Deshan Yang, 2017", in which the calculation of image entropy processed is modified by the following relation: where: h(I t ) represents the modified entropy of the processed image I t
- U represents the set of intensity levels of the digital image
- p u replaces the probability that a pixel of the digital image has the intensity u by the following relation:
- H(u) represents the level of the modified histogram for the intensity u according to the following relationship:
- a computer program product comprising program code instructions recorded on a medium readable by a computer, to implement the steps of the method as described above. , when said program is run on a computer.
- FIG.1 schematically illustrates a radiological image processing method, implemented by computer, according to one aspect of the invention
- FIG. 1 is illustrated, according to one aspect of the invention, a method of processing a radiological image / (x, y), in digital format comprising at least one radiological anomaly detected using a neural network convolution having been trained for detection of radiological anomalies on radiological examinations 1.
- the radiological image /(x,y) is characterized by the intensity of each of its pixels /(x,y), and by at least one radiological anomaly influence map C k (x,y) 2 which assigns, for each pixel (x,y) of the radiological image /(x,y) , a value representative of the proportion in which the pixel has had an influence in the result of detection of the radiological anomaly k.
- the method is implemented by computer, and comprises the steps of:
- a method according to the invention allows automatic processing of radiological images /(x,y) which improves the visibility of the regions in which the presence of an anomaly is detected. This invention therefore facilitates the diagnosis of pathologies from radiological examinations without the intervention of a radiologist or another user, and provides an image on which the visibility of possible radiological anomalies is improved, facilitating the study by a radiologist .
- the existing processing operations are modified to increase the influence of the areas of the image on which an anomaly has been detected.
- the treatments will be more effective on these regions, facilitating the highlighting of the anomalies detected.
- the detection step 1 can use a convolutional neural network having been trained for the detection of radiological anomalies representative of pathologies on radiological examinations.
- This type of network often requires the input image to have a specific resolution and number of channels (one for a gray scale image, three for an RGB image). If the input image is of a different resolution, the image can be scaled to the requested resolution and the number of channels adapted.
- Honolulu, HI IEEE. https://doi.org/10.1109/CVPR.2017.243) which takes as input images of size 224x224 with three channels.
- the radiological images of arbitrary resolution and with a single channel, are resized to 224x224 and converted into three channels, repeating the same value for each channel.
- the influence estimation step 2 can be implemented according to state-of-the-art estimates, for example, as proposed in the document "Grad-cam: Visual explanations from deep networks via gradient- based localization" Proceedings of the IEEE international conference on computer vision, 618-626, 2017, by Selvaraju, Ramprasaath R, Michael Cogswell, Abhishek Das, Ramakrishna Vedantam, Devi Parikh, and Dhruv Batra, compatible with a multitude of different architectures, which produces a map with a resolution equal to the resolution of the output of a chosen convolution layer in the sensing network architecture.
- the radiological anomaly influence map must then be resized, for example by bilinear interpolation, to the original resolution of the image.
- C k (x,y) the radiological anomaly influence map corresponding to the detected anomaly k.
- Any other radiological anomaly influence map estimate 2 compatible with the chosen detection method may be valid.
- the radiological anomaly influence maps C k (x,y) can be normalized in different ways.
- An implementation solution for the merging step 4 of the normalized radiological anomaly influence maps C kn (x,y) of each detected anomaly k is to average the maps weighted by the probability of presence of each
- represents the number of anomalies detected
- p k represents the probability p k of the presence of each anomaly calculated by the convolutional neural network.
- the cards can be merged in different ways.
- the processing step can use an intensity histogram equalization, the calculation of the histogram of which is modified by weighting the contribution of each pixel according to the following relationship: in which:
- /(x,y) represents the intensity of the pixel (x,y).
- the processing can use a variation of the method proposed in the document “Automatic x-ray image contrast enhancement based on parameter auto-optimization”. Journal of Applied Clinical Medical Physics 18(6):218-23, by Qiu, Jianfeng, H. Harold Li, Tiezhi Zhang, Fangfang Ma, and Deshan Yang, 2017", in which the calculation of image entropy processed is modified by the following relation: log Pu where: h(I t ) represents the modified entropy of the processed image I t
- U represents the set of intensity levels of the digital image
- p u replaces the probability that a pixel of the digital image has the intensity u by the following relation:
- H(u) represents the level of the modified histogram for the intensity u according to the following relationship:
- the processing 5 can also be applied differently to the processing methods mentioned above as well as to other known processing methods.
- the present invention can be implemented on a computer program product comprising computer code executable by computer, stored on a computer-readable medium and adapted to implement the method as previously described.
- the invention can be implemented on a local computer or on a network distributed platform.
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR2010735A FR3115385B1 (fr) | 2020-10-20 | 2020-10-20 | Procédé de traitement d'image radiologique |
| PCT/EP2021/078760 WO2022084223A1 (fr) | 2020-10-20 | 2021-10-18 | Procede de traitement d'image radiologique |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4232998A1 true EP4232998A1 (fr) | 2023-08-30 |
Family
ID=74553933
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP21794159.0A Pending EP4232998A1 (fr) | 2020-10-20 | 2021-10-18 | Procede de traitement d'image radiologique |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US12536653B2 (fr) |
| EP (1) | EP4232998A1 (fr) |
| JP (1) | JP2023548796A (fr) |
| CN (1) | CN116528766B (fr) |
| FR (1) | FR3115385B1 (fr) |
| WO (1) | WO2022084223A1 (fr) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117934460B (zh) * | 2024-03-21 | 2024-06-25 | 深圳市金利源绝缘材料有限公司 | 基于视觉检测的绝缘板表面缺陷智能检测方法 |
| CN119880192B (zh) * | 2025-01-15 | 2026-01-16 | 北京航空航天大学 | 一种融合电学成像和单路吸收光谱的火焰温度成像测量方法 |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100562290C (zh) * | 2005-05-31 | 2009-11-25 | 柯尼卡美能达医疗印刷器材株式会社 | 图像处理方法以及图像处理装置 |
| RU2431196C1 (ru) * | 2010-03-31 | 2011-10-10 | Закрытое Акционерное Общество "Импульс" | Способ определения уровня яркости в зоне интереса цифрового медицинского рентгеновского изображения |
| US8861886B2 (en) * | 2011-04-14 | 2014-10-14 | Carestream Health, Inc. | Enhanced visualization for medical images |
| CN108447040A (zh) * | 2018-02-09 | 2018-08-24 | 深圳市朗驰欣创科技股份有限公司 | 直方图均衡化方法、装置及终端设备 |
| US10909671B2 (en) * | 2018-10-02 | 2021-02-02 | International Business Machines Corporation | Region of interest weighted anomaly detection |
| WO2020110776A1 (fr) * | 2018-11-28 | 2020-06-04 | 富士フイルム株式会社 | Dispositif de classification, programme et procédé de classification, et dispositif d'affichage de résultat de classification |
| JP7317498B2 (ja) * | 2018-12-14 | 2023-07-31 | キヤノン株式会社 | 処理システム、処理装置、処理方法、およびプログラム |
| JP7302368B2 (ja) * | 2019-08-20 | 2023-07-04 | コニカミノルタ株式会社 | 医用情報処理装置及びプログラム |
| JP6815711B1 (ja) * | 2020-01-31 | 2021-01-20 | 学校法人慶應義塾 | 診断支援プログラム、装置、及び方法 |
| US11449717B2 (en) * | 2020-03-12 | 2022-09-20 | Fujifilm Business Innovation Corp. | System and method for identification and localization of images using triplet loss and predicted regions |
| US12056880B2 (en) * | 2020-08-03 | 2024-08-06 | Korea Advanced Institute Of Science And Technology | Method of classifying lesion of chest x-ray radiograph based on data normalization and local patch and apparatus thereof |
-
2020
- 2020-10-20 FR FR2010735A patent/FR3115385B1/fr active Active
-
2021
- 2021-10-18 EP EP21794159.0A patent/EP4232998A1/fr active Pending
- 2021-10-18 US US18/032,551 patent/US12536653B2/en active Active
- 2021-10-18 JP JP2023524133A patent/JP2023548796A/ja active Pending
- 2021-10-18 CN CN202180077690.9A patent/CN116528766B/zh active Active
- 2021-10-18 WO PCT/EP2021/078760 patent/WO2022084223A1/fr not_active Ceased
Also Published As
| Publication number | Publication date |
|---|---|
| JP2023548796A (ja) | 2023-11-21 |
| FR3115385A1 (fr) | 2022-04-22 |
| FR3115385B1 (fr) | 2023-03-24 |
| WO2022084223A1 (fr) | 2022-04-28 |
| US20230289958A1 (en) | 2023-09-14 |
| CN116528766A (zh) | 2023-08-01 |
| US12536653B2 (en) | 2026-01-27 |
| CN116528766B (zh) | 2026-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12165292B2 (en) | Generating an image mask for a digital image by utilizing a multi-branch masking pipeline with neural networks | |
| US11393100B2 (en) | Automatically generating a trimap segmentation for a digital image by utilizing a trimap generation neural network | |
| CN109410219B (zh) | 一种基于金字塔融合学习的图像分割方法、装置和计算机可读存储介质 | |
| EP0949588A1 (fr) | Procédé de recalage de deux images differentes d'un meme objet | |
| CN117115184B (zh) | 一种医疗图像分割模型的训练方法、分割方法及相关产品 | |
| EP0945830A1 (fr) | Procédé de traitement d'images incluant des étapes de segmentation d'une image multidimentionnelle et appareil d'imagerie médicale utilisant ce procédé | |
| CN113724267B (zh) | 一种乳腺超声图像肿瘤分割方法及装置 | |
| WO2022084223A1 (fr) | Procede de traitement d'image radiologique | |
| Saravanan et al. | RETRACTED ARTICLE: A brain tumor image segmentation technique in image processing using ICA-LDA algorithm with ARHE model | |
| EP2435985A1 (fr) | Procede de quantification de l'evolution de pathologies impliquant des changements de volumes de corps, notamment de tumeurs | |
| FR2884013A1 (fr) | Systeme et procede d'amelioration d'images volumetriques | |
| CN119151969B (zh) | 基于扩散模型的多层次特征融合医学图像分割方法及系统 | |
| CN117635426A (zh) | 一种基于多尺度注意力的磁共振图像超分辨率重建方法 | |
| Zhang et al. | Photo-realistic dehazing via contextual generative adversarial networks | |
| CN118447041A (zh) | 肿瘤组织细胞图像分割方法、装置、设备、介质及产品 | |
| Joshua et al. | An adaptive low-light image enhancement method via fusion of a new intuitionistic fuzzy generator and fractal-fractional derivative | |
| Dogra et al. | Effective image fusion strategies in scientific signal processing disciplines: Application to cancer and carcinoma treatment planning | |
| Kala et al. | Brain magnetic resonance image inpainting via deep edge region-based generative adversarial network | |
| Vijila Rani et al. | Detection of cervix tumor using an intelligent system accompanied with PNN classification approach | |
| CN120318072A (zh) | 图像超分辨率重建方法、装置、设备、介质及程序产品 | |
| CN111507950B (zh) | 图像分割的方法和装置、电子设备及计算机可读存储介质 | |
| Hu et al. | A novel CNN architecture for image restoration with implicit frequency selection | |
| Ayoub et al. | High quality dehazed image and video based on enhanced multi-scale guided filtering dehazing technique | |
| Nanthini et al. | A DL-DeNoiseNet Framework for Robust Image Denoising and Artifact Reduction in Medical Image Diagnostics | |
| Li et al. | Multi-scale error-driven dense residual network for image super-resolution reconstruction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20230418 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
| 17Q | First examination report despatched |
Effective date: 20251223 |