Saved in:
Bibliographic Details
Main Authors: Huo, Guohao, Dai, Ruiting, Liu, Jinliang, Shao, Ling, Tang, Hao
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.03829
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918137828474880
author Huo, Guohao
Dai, Ruiting
Liu, Jinliang
Shao, Ling
Tang, Hao
author_facet Huo, Guohao
Dai, Ruiting
Liu, Jinliang
Shao, Ling
Tang, Hao
contents In deep-sea exploration and surgical robotics scenarios, environmental lighting and device resolution limitations often cause high-frequency feature attenuation. Addressing the differences in frequency band sensitivity between CNNs and the human visual system (mid-frequency sensitivity with low-frequency sensitivity surpassing high-frequency), we experimentally quantified the CNN contrast sensitivity function and proposed a wavelet adaptive spectrum fusion (WASF) method inspired by biological vision mechanisms to balance cross-frequency image features. Furthermore, we designed a perception frequency block (PFB) that integrates WASF to enhance frequency-domain feature extraction. Based on this, we developed the FE-UNet model, which employs a SAM2 backbone network and incorporates fine-tuned Hiera-Large modules to ensure segmentation accuracy while improving generalization capability. Experiments demonstrate that FE-UNet achieves state-of-the-art performance in cross-domain tasks such as marine organism segmentation and polyp segmentation, showcasing robust adaptability and significant application potential. The code will be released soon.
format Preprint
id arxiv_https___arxiv_org_abs_2502_03829
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Frequency Domain Enhanced U-Net for Low-Frequency Information-Rich Image Segmentation in Surgical and Deep-Sea Exploration Robots
Huo, Guohao
Dai, Ruiting
Liu, Jinliang
Shao, Ling
Tang, Hao
Computer Vision and Pattern Recognition
In deep-sea exploration and surgical robotics scenarios, environmental lighting and device resolution limitations often cause high-frequency feature attenuation. Addressing the differences in frequency band sensitivity between CNNs and the human visual system (mid-frequency sensitivity with low-frequency sensitivity surpassing high-frequency), we experimentally quantified the CNN contrast sensitivity function and proposed a wavelet adaptive spectrum fusion (WASF) method inspired by biological vision mechanisms to balance cross-frequency image features. Furthermore, we designed a perception frequency block (PFB) that integrates WASF to enhance frequency-domain feature extraction. Based on this, we developed the FE-UNet model, which employs a SAM2 backbone network and incorporates fine-tuned Hiera-Large modules to ensure segmentation accuracy while improving generalization capability. Experiments demonstrate that FE-UNet achieves state-of-the-art performance in cross-domain tasks such as marine organism segmentation and polyp segmentation, showcasing robust adaptability and significant application potential. The code will be released soon.
title Frequency Domain Enhanced U-Net for Low-Frequency Information-Rich Image Segmentation in Surgical and Deep-Sea Exploration Robots
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.03829