FALCON: Frequency Adjoint Link with CONtinuous Density Mask for Fast Single Image Dehazing

TL;DR

FALCON is a fast and high-quality single-image dehazing system that uses frequency space processing and a continuous density mask to achieve real-time performance suitable for time-sensitive applications.

Contribution

The paper introduces FALCON, a novel dehazing network with a frequency domain bottleneck module and a continuous density mask, enabling state-of-the-art speed and quality.

Findings

Achieves over 180 frames-per-second in dehazing tasks.

Outperforms existing methods in PSNR and SSIM metrics.

Demonstrates effectiveness through comprehensive experiments and ablation studies.

Abstract

Image dehazing, addressing atmospheric interference like fog and haze, remains a pervasive challenge crucial for robust vision applications such as surveillance and remote sensing under adverse visibility. While various methodologies have evolved from early works predicting transmission matrix and atmospheric light features to deep learning and dehazing networks, they innately prioritize dehazing quality metrics, neglecting the need for real-time applicability in time-sensitive domains like autonomous driving. This work introduces FALCON (Frequency Adjoint Link with CONtinuous density mask), a single-image dehazing system achieving state-of-the-art performance on both quality and speed. Particularly, we develop a novel bottleneck module, namely, Frequency Adjoint Link, operating in the frequency space to globally expand the receptive field with minimal growth in network size. Further, we leverage the underlying haze distribution based on the atmospheric scattering model via a Continuous Density Mask (CDM) which serves as a continuous-valued mask input prior and a differentiable auxiliary loss. Comprehensive experiments involving multiple state-of-the-art methods and ablation analysis demonstrate FALCON's exceptional performance in both dehazing quality and speed (i.e., >$180 frames-per-second), quantified by metrics such as FPS, PSNR, and SSIM.