dorsal/arxiv
View SchemaTranslating Light-Sheet Microscopy Images to Virtual H&E Using CycleGAN
| Authors | Yanhua Zhao |
|---|---|
| Categories | |
| ArXiv ID | 2601.08776vv1 |
| URL | https://arxiv.org/abs/2601.08776 |
| License | http://creativecommons.org/licenses/by/4.0/ |
Abstract
Histopathology analysis relies on Hematoxylin and Eosin (H&E) staining, but fluorescence microscopy offers complementary information. Converting fluorescence images to H&E-like appearance can aid interpretation and integration with standard workflows. We present a Cycle-Consistent Adversarial Network (CycleGAN) approach for unpaired image-to-image translation from multi-channel fluorescence microscopy to pseudo H&E stained histopathology images. The method combines C01 and C02 fluorescence channels into RGB and learns a bidirectional mapping between fluorescence and H&E domains without paired training data. The architecture uses ResNet-based generators with residual blocks and PatchGAN discriminators, trained with adversarial, cycle-consistency, and identity losses. Experiments on fluorescence microscopy datasets show the model generates realistic pseudo H&E images that preserve morphological structures while adopting H&E-like color characteristics. This enables visualization of fluorescence data in a format familiar to pathologists and supports integration with existing H&E-based analysis pipelines.
{
"annotation_id": "1101af7f-24f3-4733-9565-fa2a3bd264d9",
"date_created": "2026-02-17T05:53:15.716000Z",
"date_modified": "2026-02-17T05:53:15.716000Z",
"file_hash": "c926c4b7e196b54345a96173b310937e53421d6d73de9e51442f756f44ec8779",
"private": false,
"record": {
"abstract": "Histopathology analysis relies on Hematoxylin and Eosin (H\u0026E) staining, but fluorescence microscopy offers complementary information. Converting fluorescence images to H\u0026E-like appearance can aid interpretation and integration with standard workflows. We present a Cycle-Consistent Adversarial Network (CycleGAN) approach for unpaired image-to-image translation from multi-channel fluorescence microscopy to pseudo H\u0026E stained histopathology images. The method combines C01 and C02 fluorescence channels into RGB and learns a bidirectional mapping between fluorescence and H\u0026E domains without paired training data. The architecture uses ResNet-based generators with residual blocks and PatchGAN discriminators, trained with adversarial, cycle-consistency, and identity losses. Experiments on fluorescence microscopy datasets show the model generates realistic pseudo H\u0026E images that preserve morphological structures while adopting H\u0026E-like color characteristics. This enables visualization of fluorescence data in a format familiar to pathologists and supports integration with existing H\u0026E-based analysis pipelines.",
"arxiv_id": "2601.08776",
"authors": [
"Yanhua Zhao"
],
"categories": [
"cs.CV",
"cs.AI"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"title": "Translating Light-Sheet Microscopy Images to Virtual H\u0026E Using CycleGAN",
"url": "https://arxiv.org/abs/2601.08776",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "d066b4bf-5f6c-4223-8669-4f10bf319d8c",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}