PMID- 38061155 OWN - NLM STAT- MEDLINE DCOM- 20240110 LR - 20240206 IS - 1879-0534 (Electronic) IS - 0010-4825 (Linking) VI - 168 DP - 2024 Jan TI - PATrans: Pixel-Adaptive Transformer for edge segmentation of cervical nuclei on small-scale datasets. PG - 107823 LID - S0010-4825(23)01288-X [pii] LID - 10.1016/j.compbiomed.2023.107823 [doi] AB - Transformer has shown excellent performance in various visual tasks, making its application in medicine an inevitable trend. Nevertheless, simply using transformer for small-scale cervical nuclei datasets will result in disastrous performance. Scarce nuclei pixels are not enough to compensate for the lack of CNNs-inherent intrinsic inductive biases, making transformer difficult to model local visual structures and deal with scale variations. Thus, we propose a Pixel Adaptive Transformer(PATrans) to improve the segmentation performance of nuclei edges on small datasets through adaptive pixel tuning. Specifically, to mitigate information loss resulting from mapping different patches into similar latent representations, Consecutive Pixel Patch (CPP) embeds rich multi-scale context into isolated image patches. In this way, it can provide intrinsic scale invariance for 1D input sequences to maintain semantic consistency, allowing the PATrans to establish long-range dependencies quickly. Futhermore, due to the existing handcrafted-attention is agnostic to the widely varying pixel distributions, the Pixel Adaptive Transformer Block (PATB) effectively models the relationships between different pixels across the entire feature map in a data-dependent manner, guided by the important regions. By collaboratively learning local features and global dependencies, PATrans can adaptively reduce the interference of irrelevant pixels. Extensive experiments demonstrate the superiority of our model on three datasets(Ours, ISBI, Herlev). CI - Copyright (c) 2023 Elsevier Ltd. All rights reserved. FAU - Hu, Hexuan AU - Hu H AD - Key Laboratory of Water Big Data Technology of Ministry of Water Resources, Hohai University, Nanjing, 211100, PR China; College of Computer and Information, Hohai University, Nanjing, 211100, PR China. Electronic address: hexuan_hu@hhu.edu.cn. FAU - Zhang, Jianyu AU - Zhang J AD - Key Laboratory of Water Big Data Technology of Ministry of Water Resources, Hohai University, Nanjing, 211100, PR China; College of Computer and Information, Hohai University, Nanjing, 211100, PR China. Electronic address: qieyi@hhu.edu.cn. FAU - Yang, Tianjin AU - Yang T AD - Key Laboratory of Water Big Data Technology of Ministry of Water Resources, Hohai University, Nanjing, 211100, PR China; College of Computer and Information, Hohai University, Nanjing, 211100, PR China. Electronic address: yangtianjin@hhu.edu.cn. FAU - Hu, Qiang AU - Hu Q AD - Key Laboratory of Water Big Data Technology of Ministry of Water Resources, Hohai University, Nanjing, 211100, PR China; College of Computer and Information, Hohai University, Nanjing, 211100, PR China. Electronic address: huqiang@hhu.edu.cn. FAU - Yu, Yufeng AU - Yu Y AD - Key Laboratory of Water Big Data Technology of Ministry of Water Resources, Hohai University, Nanjing, 211100, PR China; College of Computer and Information, Hohai University, Nanjing, 211100, PR China. Electronic address: yfyu@hhu.edu.cn. FAU - Huang, Qian AU - Huang Q AD - Key Laboratory of Water Big Data Technology of Ministry of Water Resources, Hohai University, Nanjing, 211100, PR China; College of Computer and Information, Hohai University, Nanjing, 211100, PR China. Electronic address: huangqian@hhu.edu.cn. LA - eng PT - Journal Article DEP - 20231205 PL - United States TA - Comput Biol Med JT - Computers in biology and medicine JID - 1250250 SB - IM MH - *Cell Nucleus MH - Learning MH - *Medicine MH - Semantics MH - Image Processing, Computer-Assisted OTO - NOTNLM OT - Adherent cells and nuclei OT - Cervical cancer OT - Nuclei edge segmentation OT - Nuclei segmentation OT - Transformer COIS- Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. EDAT- 2023/12/08 00:42 MHDA- 2024/01/10 06:42 CRDT- 2023/12/07 18:02 PHST- 2023/09/13 00:00 [received] PHST- 2023/11/22 00:00 [revised] PHST- 2023/12/04 00:00 [accepted] PHST- 2024/01/10 06:42 [medline] PHST- 2023/12/08 00:42 [pubmed] PHST- 2023/12/07 18:02 [entrez] AID - S0010-4825(23)01288-X [pii] AID - 10.1016/j.compbiomed.2023.107823 [doi] PST - ppublish SO - Comput Biol Med. 2024 Jan;168:107823. doi: 10.1016/j.compbiomed.2023.107823. Epub 2023 Dec 5.