A semantic-driven frequency-consistent underwater image enhancement method

By calculating the low-frequency residual components and semantic segmentation of underwater images, and combining adaptive weight optimization, the problems of global structural distortion and local detail imbalance in underwater image enhancement are solved, achieving a balanced enhancement effect between global and local aspects and improving the overall quality of underwater images.

CN121961897BActive Publication Date: 2026-07-03WUHAN UNIV OF SCI & TECH +1
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
WUHAN UNIV OF SCI & TECH
Filing Date
2026-03-31
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Existing technologies do not fully utilize frequency domain compensation mechanisms in underwater image enhancement, making it difficult to quantify the degree of global degradation. This results in distortion of the global image structure and imbalance of local details, failing to effectively solve the imbalance between global and local enhancement.

Method used

By acquiring the low-frequency components of the underwater degraded image and the clear image, the difference of the deep feature vector is calculated to obtain the global low-frequency residual component. Combined with semantic segmentation and adaptive weight optimization, frequency domain compensation and local detail enhancement are achieved. Enhancement weights are dynamically allocated to balance global structure and local details.

Benefits of technology

It effectively improves the global structural consistency and local detail enhancement of underwater images, solves the problem of imbalance between global and local enhancement, and enhances the integrity and consistency of underwater image enhancement.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121961897B_ABST
    Figure CN121961897B_ABST
Patent Text Reader

Abstract

This invention provides a semantically driven, frequency-consistent underwater image enhancement method, belonging to the field of underwater image processing technology. The specific steps include: acquiring both degraded and clear underwater images of the same scene; extracting the low-frequency components of both; calculating and obtaining the global low-frequency residual components; generating a semantic partition mask through semantic segmentation; dividing regions by combining gray-level variance to obtain the final degraded region; introducing semantic adaptive coefficients to output semantic enhancement features; combining the semantic adaptive coefficients and gray-level variance to obtain fusion weight coefficients; weightedly fusing the semantic enhancement features with the local low-frequency residual components to obtain fusion features; inputting the combined feature map into an image enhancement network model; outputting clear underwater image fragments; and stitching them with the original image of the non-degraded region to obtain a complete underwater enhanced image. This invention achieves accurate quantification and frequency domain compensation of global degradation differences, solves the problem of global and local enhancement imbalance, and improves the integrity and consistency of underwater image enhancement.
Need to check novelty before this filing date? Find Prior Art