Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2511.12006 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908654788149248 |
|---|---|
| author | Yang, Kai-Wen K. Bai, Andrew Bermudez, Alexandra Hong, Yunqi Latham, Zoe Sloan, Iris Liu, Michael Goyal, Vishrut Hsieh, Cho-Jui Lin, Neil Y. C. |
| author_facet | Yang, Kai-Wen K. Bai, Andrew Bermudez, Alexandra Hong, Yunqi Latham, Zoe Sloan, Iris Liu, Michael Goyal, Vishrut Hsieh, Cho-Jui Lin, Neil Y. C. |
| contents | Deep learning is transforming microscopy, yet models often fail when applied to images from new instruments or acquisition settings. Conventional adversarial domain adaptation (ADDA) retrains entire networks, often disrupting learned semantic representations. Here, we overturn this paradigm by showing that adapting only the earliest convolutional layers, while freezing deeper layers, yields reliable transfer. Building on this principle, we introduce Subnetwork Image Translation ADDA with automatic depth selection (SIT-ADDA-Auto), a self-configuring framework that integrates shallow-layer adversarial alignment with predictive uncertainty to automatically select adaptation depth without target labels. We demonstrate robustness via multi-metric evaluation, blinded expert assessment, and uncertainty-depth ablations. Across exposure and illumination shifts, cross-instrument transfer, and multiple stains, SIT-ADDA improves reconstruction and downstream segmentation over full-encoder adaptation and non-adversarial baselines, with reduced drift of semantic features. Our results provide a design rule for label-free adaptation in microscopy and a recipe for field settings; the code is publicly available. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_12006 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Uncertainty-Guided Selective Adaptation Enables Cross-Platform Predictive Fluorescence Microscopy Yang, Kai-Wen K. Bai, Andrew Bermudez, Alexandra Hong, Yunqi Latham, Zoe Sloan, Iris Liu, Michael Goyal, Vishrut Hsieh, Cho-Jui Lin, Neil Y. C. Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning Deep learning is transforming microscopy, yet models often fail when applied to images from new instruments or acquisition settings. Conventional adversarial domain adaptation (ADDA) retrains entire networks, often disrupting learned semantic representations. Here, we overturn this paradigm by showing that adapting only the earliest convolutional layers, while freezing deeper layers, yields reliable transfer. Building on this principle, we introduce Subnetwork Image Translation ADDA with automatic depth selection (SIT-ADDA-Auto), a self-configuring framework that integrates shallow-layer adversarial alignment with predictive uncertainty to automatically select adaptation depth without target labels. We demonstrate robustness via multi-metric evaluation, blinded expert assessment, and uncertainty-depth ablations. Across exposure and illumination shifts, cross-instrument transfer, and multiple stains, SIT-ADDA improves reconstruction and downstream segmentation over full-encoder adaptation and non-adversarial baselines, with reduced drift of semantic features. Our results provide a design rule for label-free adaptation in microscopy and a recipe for field settings; the code is publicly available. |
| title | Uncertainty-Guided Selective Adaptation Enables Cross-Platform Predictive Fluorescence Microscopy |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2511.12006 |