Systematic Augmentation in HSV Space for Semantic Segmentation of Prostate Biopsies
(2023) 22nd Scandinavian Conference on Image Analysis, SCIA 2023 In Lecture Notes in Computer Science 13886. p.293-308- Abstract
- In recent years, the combination of the digitization of the field of pathology and increased computational power has led to a big increase in research of computer-aided diagnostics using systems based on artificial intelligence (AI). This includes detection and classification of prostate cancer, where several studies have shown great promise in automated prostate cancer grading using deep learning based AI systems. However, there is still work to be done to ensure that these algorithms are invariant to possible variations of the digitized microscopy images they are applied to. A standard method in deep learning to increase the variation of the training data is dataset augmentation. All of these studies apply some augmentation of their... (More)
- In recent years, the combination of the digitization of the field of pathology and increased computational power has led to a big increase in research of computer-aided diagnostics using systems based on artificial intelligence (AI). This includes detection and classification of prostate cancer, where several studies have shown great promise in automated prostate cancer grading using deep learning based AI systems. However, there is still work to be done to ensure that these algorithms are invariant to possible variations of the digitized microscopy images they are applied to. A standard method in deep learning to increase the variation of the training data is dataset augmentation. All of these studies apply some augmentation of their data, however, there is a lack of evaluation of different methods and their impact on this crucial part of the AI systems. In this study, we look into different color augmentation methods for the task of segmentation of prostate biopsies. Furthermore, we introduce a novel color augmentation method based on stereographic projection. Our results affirm the importance of studying different augmentation methods and indicate a gain in performance using our method. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/eb0928b5-b2fc-4209-b1b8-da4d53eab2ab
- author
- Winzell, Filip LU ; Arvidsson, Ida LU ; Overgaard, Niels Christian LU ; Åström, Kalle LU ; Marginean, Felicia-Elena LU ; Bjartell, Anders LU ; Krzyzanowska, Agnieszka LU ; Simoulis, Athanasios LU and Heyden, Anders LU
- organization
-
- Mathematics (Faculty of Engineering)
- LTH Profile Area: AI and Digitalization
- eSSENCE: The e-Science Collaboration
- ELLIIT: the Linköping-Lund initiative on IT and mobile communication
- LTH Profile Area: Engineering Health
- Engineering Mathematics (M.Sc.Eng.)
- Mathematical Imaging Group (research group)
- Partial differential equations (research group)
- LU Profile Area: Light and Materials
- LU Profile Area: Natural and Artificial Cognition
- Stroke Imaging Research group (research group)
- Division of Translational Cancer Research
- LUCC: Lund University Cancer Centre
- Urological cancer, Malmö (research group)
- EpiHealth: Epidemiology for Health
- Centre for Mathematical Sciences
- publishing date
- 2023-04-26
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Image Analysis : 23rd Scandinavian Conference, SCIA 2023, Proceedings, Part II - 23rd Scandinavian Conference, SCIA 2023, Proceedings, Part II
- series title
- Lecture Notes in Computer Science
- volume
- 13886
- pages
- 293 - 308
- publisher
- Springer
- conference name
- 22nd Scandinavian Conference on Image Analysis, SCIA 2023
- conference location
- Sirkka, Finland
- conference dates
- 2023-04-18 - 2023-04-21
- external identifiers
-
- scopus:85161427190
- ISSN
- 0302-9743
- 1611-3349
- ISBN
- 978-3-031-31438-4
- 978-3-031-31437-7
- DOI
- 10.1007/978-3-031-31438-4_20
- language
- English
- LU publication?
- yes
- id
- eb0928b5-b2fc-4209-b1b8-da4d53eab2ab
- date added to LUP
- 2023-04-27 14:33:11
- date last changed
- 2024-09-07 10:19:10
@inproceedings{eb0928b5-b2fc-4209-b1b8-da4d53eab2ab, abstract = {{In recent years, the combination of the digitization of the field of pathology and increased computational power has led to a big increase in research of computer-aided diagnostics using systems based on artificial intelligence (AI). This includes detection and classification of prostate cancer, where several studies have shown great promise in automated prostate cancer grading using deep learning based AI systems. However, there is still work to be done to ensure that these algorithms are invariant to possible variations of the digitized microscopy images they are applied to. A standard method in deep learning to increase the variation of the training data is dataset augmentation. All of these studies apply some augmentation of their data, however, there is a lack of evaluation of different methods and their impact on this crucial part of the AI systems. In this study, we look into different color augmentation methods for the task of segmentation of prostate biopsies. Furthermore, we introduce a novel color augmentation method based on stereographic projection. Our results affirm the importance of studying different augmentation methods and indicate a gain in performance using our method.}}, author = {{Winzell, Filip and Arvidsson, Ida and Overgaard, Niels Christian and Åström, Kalle and Marginean, Felicia-Elena and Bjartell, Anders and Krzyzanowska, Agnieszka and Simoulis, Athanasios and Heyden, Anders}}, booktitle = {{Image Analysis : 23rd Scandinavian Conference, SCIA 2023, Proceedings, Part II}}, isbn = {{978-3-031-31438-4}}, issn = {{0302-9743}}, language = {{eng}}, month = {{04}}, pages = {{293--308}}, publisher = {{Springer}}, series = {{Lecture Notes in Computer Science}}, title = {{Systematic Augmentation in HSV Space for Semantic Segmentation of Prostate Biopsies}}, url = {{http://dx.doi.org/10.1007/978-3-031-31438-4_20}}, doi = {{10.1007/978-3-031-31438-4_20}}, volume = {{13886}}, year = {{2023}}, }