Cargando…

RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis

We propose a novel transfer learning framework for pathological image analysis, the Response-based Cross-task Knowledge Distillation (RCKD), which improves the performance of the model by pretraining it on a large unlabeled dataset guided by a high-performance teacher model. RCKD first pretrains a s...

Descripción completa

Detalles Bibliográficos
Autores principales: Kim, Hyunil, Kwak, Tae-Yeong, Chang, Hyeyoon, Kim, Sun Woo, Kim, Injung
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10669242/
https://www.ncbi.nlm.nih.gov/pubmed/38002403
http://dx.doi.org/10.3390/bioengineering10111279
_version_ 1785139649718517760
author Kim, Hyunil
Kwak, Tae-Yeong
Chang, Hyeyoon
Kim, Sun Woo
Kim, Injung
author_facet Kim, Hyunil
Kwak, Tae-Yeong
Chang, Hyeyoon
Kim, Sun Woo
Kim, Injung
author_sort Kim, Hyunil
collection PubMed
description We propose a novel transfer learning framework for pathological image analysis, the Response-based Cross-task Knowledge Distillation (RCKD), which improves the performance of the model by pretraining it on a large unlabeled dataset guided by a high-performance teacher model. RCKD first pretrains a student model to predict the nuclei segmentation results of the teacher model for unlabeled pathological images, and then fine-tunes the pretrained model for the downstream tasks, such as organ cancer sub-type classification and cancer region segmentation, using relatively small target datasets. Unlike conventional knowledge distillation, RCKD does not require that the target tasks of the teacher and student models be the same. Moreover, unlike conventional transfer learning, RCKD can transfer knowledge between models with different architectures. In addition, we propose a lightweight architecture, the Convolutional neural network with Spatial Attention by Transformers (CSAT), for processing high-resolution pathological images with limited memory and computation. CSAT exhibited a top-1 accuracy of 78.6% on ImageNet with only 3M parameters and 1.08 G multiply-accumulate (MAC) operations. When pretrained by RCKD, CSAT exhibited average classification and segmentation accuracies of 94.2% and 0.673 mIoU on six pathological image datasets, which is 4% and 0.043 mIoU higher than EfficientNet-B0, and 7.4% and 0.006 mIoU higher than ConvNextV2-Atto pretrained on ImageNet, respectively.
format Online
Article
Text
id pubmed-10669242
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-106692422023-11-02 RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis Kim, Hyunil Kwak, Tae-Yeong Chang, Hyeyoon Kim, Sun Woo Kim, Injung Bioengineering (Basel) Article We propose a novel transfer learning framework for pathological image analysis, the Response-based Cross-task Knowledge Distillation (RCKD), which improves the performance of the model by pretraining it on a large unlabeled dataset guided by a high-performance teacher model. RCKD first pretrains a student model to predict the nuclei segmentation results of the teacher model for unlabeled pathological images, and then fine-tunes the pretrained model for the downstream tasks, such as organ cancer sub-type classification and cancer region segmentation, using relatively small target datasets. Unlike conventional knowledge distillation, RCKD does not require that the target tasks of the teacher and student models be the same. Moreover, unlike conventional transfer learning, RCKD can transfer knowledge between models with different architectures. In addition, we propose a lightweight architecture, the Convolutional neural network with Spatial Attention by Transformers (CSAT), for processing high-resolution pathological images with limited memory and computation. CSAT exhibited a top-1 accuracy of 78.6% on ImageNet with only 3M parameters and 1.08 G multiply-accumulate (MAC) operations. When pretrained by RCKD, CSAT exhibited average classification and segmentation accuracies of 94.2% and 0.673 mIoU on six pathological image datasets, which is 4% and 0.043 mIoU higher than EfficientNet-B0, and 7.4% and 0.006 mIoU higher than ConvNextV2-Atto pretrained on ImageNet, respectively. MDPI 2023-11-02 /pmc/articles/PMC10669242/ /pubmed/38002403 http://dx.doi.org/10.3390/bioengineering10111279 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Kim, Hyunil
Kwak, Tae-Yeong
Chang, Hyeyoon
Kim, Sun Woo
Kim, Injung
RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis
title RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis
title_full RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis
title_fullStr RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis
title_full_unstemmed RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis
title_short RCKD: Response-Based Cross-Task Knowledge Distillation for Pathological Image Analysis
title_sort rckd: response-based cross-task knowledge distillation for pathological image analysis
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10669242/
https://www.ncbi.nlm.nih.gov/pubmed/38002403
http://dx.doi.org/10.3390/bioengineering10111279
work_keys_str_mv AT kimhyunil rckdresponsebasedcrosstaskknowledgedistillationforpathologicalimageanalysis
AT kwaktaeyeong rckdresponsebasedcrosstaskknowledgedistillationforpathologicalimageanalysis
AT changhyeyoon rckdresponsebasedcrosstaskknowledgedistillationforpathologicalimageanalysis
AT kimsunwoo rckdresponsebasedcrosstaskknowledgedistillationforpathologicalimageanalysis
AT kiminjung rckdresponsebasedcrosstaskknowledgedistillationforpathologicalimageanalysis