Abstract:Nuclei appear small in size, yet, in real clinical practice, the global spatial information and correlation of the color or brightness contrast between nuclei and background, have been considered a crucial component for accurate nuclei segmentation. However, the field of automatic nuclei segmentation is dominated by Convolutional Neural Networks (CNNs), meanwhile, the potential of the recently prevalent Transformers has not been fully explored, which is powerful in capturing local-global correlations. To this end, we make the first attempt at a pure Transformer framework for nuclei segmentation, called TransNuSeg. Different from prior work, we decouple the challenging nuclei segmentation task into an intrinsic multi-task learning task, where a tri-decoder structure is employed for nuclei instance, nuclei edge, and clustered edge segmentation respectively. To eliminate the divergent predictions from different branches in previous work, a novel self distillation loss is introduced to explicitly impose consistency regulation between branches. Moreover, to formulate the high correlation between branches and also reduce the number of parameters, an efficient attention sharing scheme is proposed by partially sharing the self-attention heads amongst the tri-decoders. Finally, a token MLP bottleneck replaces the over-parameterized Transformer bottleneck for a further reduction in model complexity. Experiments on two datasets of different modalities, including MoNuSeg have shown that our methods can outperform state-of-the-art counterparts such as CA2.5-Net by 2-3% Dice with 30% fewer parameters. In conclusion, TransNuSeg confirms the strength of Transformer in the context of nuclei segmentation, which thus can serve as an efficient solution for real clinical practice. Code is available at <a class="link-external link-https" href="https://github.com/zhenqi-he/transnuseg" rel="external noopener nofollow">this https URL</a>.

The NiuTrans System for WNGT 2020 Efficiency Task

OpenNMT System Description for WNMT 2018: 800 Words/sec on a Single-Core CPU

NJUNLP’s Machine Translation System for CCMT-2020 Uighur → Chinese Translation Task

The Volctrans Neural Speech Translation System for IWSLT 2021

NJUNLP’s Machine Translation System for CCMT-2020 Uighur $$\rightarrow $$ Chinese Translation Task

Baidu Neural Machine Translation Systems for WMT19.

WeChat Neural Machine Translation Systems for WMT20

Tsinghua University Neural Machine Translation Systems for CCMT 2020

The YiTrans End-to-End Speech Translation System for IWSLT 2022 Offline Shared Task

NICT's Neural and Statistical Machine Translation Systems for the WMT18 News Translation Task

DiDi's Machine Translation System for WMT2020

WeChat Neural Machine Translation Systems for WMT21

SJTU-NICT's Supervised and Unsupervised Neural Machine Translation Systems for the WMT20 News Translation Task

TransNuSeg: A Lightweight Multi-Task Transformer for Nuclei Segmentation

The Sogou-TIIC Speech Translation System for IWSLT 2018.

Towards better Chinese-centric neural machine translation for low-resource languages

XMU Neural Machine Translation Systems for WMT 17.

NVIDIA NeMo Neural Machine Translation Systems for English-German and English-Russian News and Biomedical Tasks at WMT21

LIT Team's System Description for Japanese-Chinese Machine Translation Task in IWSLT 2020.

The Volctrans GLAT System: Non-autoregressive Translation Meets WMT21