Automated classification of polyps using deep learning architectures and few-shot learning
Abstract Background Colorectal cancer is a leading cause of cancer-related deaths worldwide. The best method to prevent CRC is a colonoscopy. However, not all colon polyps have the risk of becoming cancerous. Therefore, polyps are classified using different classification systems. After the classifi...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2023-04-01
|
| Series: | BMC Medical Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12880-023-01007-4 |
| _version_ | 1797840806500892672 |
|---|---|
| author | Adrian Krenzer Stefan Heil Daniel Fitting Safa Matti Wolfram G. Zoller Alexander Hann Frank Puppe |
| author_facet | Adrian Krenzer Stefan Heil Daniel Fitting Safa Matti Wolfram G. Zoller Alexander Hann Frank Puppe |
| author_sort | Adrian Krenzer |
| collection | DOAJ |
| description | Abstract Background Colorectal cancer is a leading cause of cancer-related deaths worldwide. The best method to prevent CRC is a colonoscopy. However, not all colon polyps have the risk of becoming cancerous. Therefore, polyps are classified using different classification systems. After the classification, further treatment and procedures are based on the classification of the polyp. Nevertheless, classification is not easy. Therefore, we suggest two novel automated classifications system assisting gastroenterologists in classifying polyps based on the NICE and Paris classification. Methods We build two classification systems. One is classifying polyps based on their shape (Paris). The other classifies polyps based on their texture and surface patterns (NICE). A two-step process for the Paris classification is introduced: First, detecting and cropping the polyp on the image, and secondly, classifying the polyp based on the cropped area with a transformer network. For the NICE classification, we design a few-shot learning algorithm based on the Deep Metric Learning approach. The algorithm creates an embedding space for polyps, which allows classification from a few examples to account for the data scarcity of NICE annotated images in our database. Results For the Paris classification, we achieve an accuracy of 89.35 %, surpassing all papers in the literature and establishing a new state-of-the-art and baseline accuracy for other publications on a public data set. For the NICE classification, we achieve a competitive accuracy of 81.13 % and demonstrate thereby the viability of the few-shot learning paradigm in polyp classification in data-scarce environments. Additionally, we show different ablations of the algorithms. Finally, we further elaborate on the explainability of the system by showing heat maps of the neural network explaining neural activations. Conclusion Overall we introduce two polyp classification systems to assist gastroenterologists. We achieve state-of-the-art performance in the Paris classification and demonstrate the viability of the few-shot learning paradigm in the NICE classification, addressing the prevalent data scarcity issues faced in medical machine learning. |
| first_indexed | 2024-04-09T16:20:41Z |
| format | Article |
| id | doaj.art-aa67574667b040a08c95c4d9d717135a |
| institution | Directory Open Access Journal |
| issn | 1471-2342 |
| language | English |
| last_indexed | 2024-04-09T16:20:41Z |
| publishDate | 2023-04-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Medical Imaging |
| spelling | doaj.art-aa67574667b040a08c95c4d9d717135a2023-04-23T11:31:51ZengBMCBMC Medical Imaging1471-23422023-04-0123112510.1186/s12880-023-01007-4Automated classification of polyps using deep learning architectures and few-shot learningAdrian Krenzer0Stefan Heil1Daniel Fitting2Safa Matti3Wolfram G. Zoller4Alexander Hann5Frank Puppe6Department of Artificial Intelligence and Knowledge Systems, Julius-Maximilians University of WürzburgDepartment of Artificial Intelligence and Knowledge Systems, Julius-Maximilians University of WürzburgInterventional and Experimental Endoscopy (InExEn), Department of Internal Medicine II, University Hospital WürzburgDepartment of Artificial Intelligence and Knowledge Systems, Julius-Maximilians University of WürzburgDepartment of Internal Medicine and Gastroenterology, KatharinenhospitalInterventional and Experimental Endoscopy (InExEn), Department of Internal Medicine II, University Hospital WürzburgDepartment of Artificial Intelligence and Knowledge Systems, Julius-Maximilians University of WürzburgAbstract Background Colorectal cancer is a leading cause of cancer-related deaths worldwide. The best method to prevent CRC is a colonoscopy. However, not all colon polyps have the risk of becoming cancerous. Therefore, polyps are classified using different classification systems. After the classification, further treatment and procedures are based on the classification of the polyp. Nevertheless, classification is not easy. Therefore, we suggest two novel automated classifications system assisting gastroenterologists in classifying polyps based on the NICE and Paris classification. Methods We build two classification systems. One is classifying polyps based on their shape (Paris). The other classifies polyps based on their texture and surface patterns (NICE). A two-step process for the Paris classification is introduced: First, detecting and cropping the polyp on the image, and secondly, classifying the polyp based on the cropped area with a transformer network. For the NICE classification, we design a few-shot learning algorithm based on the Deep Metric Learning approach. The algorithm creates an embedding space for polyps, which allows classification from a few examples to account for the data scarcity of NICE annotated images in our database. Results For the Paris classification, we achieve an accuracy of 89.35 %, surpassing all papers in the literature and establishing a new state-of-the-art and baseline accuracy for other publications on a public data set. For the NICE classification, we achieve a competitive accuracy of 81.13 % and demonstrate thereby the viability of the few-shot learning paradigm in polyp classification in data-scarce environments. Additionally, we show different ablations of the algorithms. Finally, we further elaborate on the explainability of the system by showing heat maps of the neural network explaining neural activations. Conclusion Overall we introduce two polyp classification systems to assist gastroenterologists. We achieve state-of-the-art performance in the Paris classification and demonstrate the viability of the few-shot learning paradigm in the NICE classification, addressing the prevalent data scarcity issues faced in medical machine learning.https://doi.org/10.1186/s12880-023-01007-4Machine learningDeep learningEndoscopyGastroenterologyAutomationImage classification |
| spellingShingle | Adrian Krenzer Stefan Heil Daniel Fitting Safa Matti Wolfram G. Zoller Alexander Hann Frank Puppe Automated classification of polyps using deep learning architectures and few-shot learning BMC Medical Imaging Machine learning Deep learning Endoscopy Gastroenterology Automation Image classification |
| title | Automated classification of polyps using deep learning architectures and few-shot learning |
| title_full | Automated classification of polyps using deep learning architectures and few-shot learning |
| title_fullStr | Automated classification of polyps using deep learning architectures and few-shot learning |
| title_full_unstemmed | Automated classification of polyps using deep learning architectures and few-shot learning |
| title_short | Automated classification of polyps using deep learning architectures and few-shot learning |
| title_sort | automated classification of polyps using deep learning architectures and few shot learning |
| topic | Machine learning Deep learning Endoscopy Gastroenterology Automation Image classification |
| url | https://doi.org/10.1186/s12880-023-01007-4 |
| work_keys_str_mv | AT adriankrenzer automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning AT stefanheil automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning AT danielfitting automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning AT safamatti automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning AT wolframgzoller automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning AT alexanderhann automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning AT frankpuppe automatedclassificationofpolypsusingdeeplearningarchitecturesandfewshotlearning |