CLIP-Guided Data Augmentation for Night-Time Image Dehazing

• URL: http://arxiv.org/abs/2604.05500v1
• Date: Tue, 07 Apr 2026 06:43:24 GMT
• Title: CLIP-Guided Data Augmentation for Night-Time Image Dehazing
• Abstract summary: This paper presents our solution to the NTIRE 2026 Night Time Image Dehazing Challenge, built as a unified framework.<n> NAFNet is trained in two stages, first adapting to the target domain and then expanding to broader degradation patterns.<n>At inference time, TLC, x8 self-ensemble, and weighted snapshot fusion are combined to improve output stability.
• Score: 4.407514620852976
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Nighttime image dehazing faces a more complex degradation pattern than its daytime counterpart, as haze scattering couples with low illumination, non-uniform lighting, and strong light interference. Under limited supervision, this complexity aggravates domain drift and training instability, since target-domain samples are scarce while naively introducing external data may weaken adaptation due to distribution mismatch. This paper presents our solution to the NTIRE 2026 Night Time Image Dehazing Challenge, built as a unified framework that integrates domain-aligned data construction, stage-wise training, and inference-time enhancement. Specifically, a pre-trained CLIP visual encoder screens candidate external samples by similarity to construct training data closer to the target domain. NAFNet is then trained in two stages, first adapting to the target domain and then expanding to broader degradation patterns. At inference time, TLC, x8 self-ensemble, and weighted snapshot fusion are combined to improve output stability. Rather than relying on complex network redesign, the proposed framework offers a practical and effective pipeline for nighttime image dehazing.

Related papers

• UBLLIE: Unified Backlight and Low-Light Image Enhancement [15.386282490743838]
  Backlit and low-light images often suffer from severe exposure imbalance or global underexposure.<n>We propose a unified, unsupervised enhancement framework that addresses both types of degradation.<n>Our framework consistently outperforms state-of-the-art supervised and unsupervised methods in terms of fidelity, perceptual quality, and generalization.
  arXiv  Detail & Related papers  (2026-08-05T04:17:05Z)
• EvLIR: Learning Illumination Residuals from Ordered Events for Low-Light Image Enhancement [13.751247283539177]
  EvLIR learns illumination residuals from ordered events for low-light image enhancement.<n>On SDE and SDSD indoor/outdoor benchmarks, EvLIR achieves the best result on eleven of twelve dataset-metric pairs.
  arXiv  Detail & Related papers  (2026-06-28T14:47:49Z)
• HistoFusionNet: Histogram-Guided Fusion and Frequency-Adaptive Refinement for Nighttime Image Dehazing [20.95434972260448]
  HistoFusionNet is a transformer-enhanced architecture tailored for nighttime image dehazing.<n>Our team ranked 1st among 22 participating teams, highlighting the robustness and competitive performance of our method.
  arXiv  Detail & Related papers  (2026-04-04T17:08:13Z)
• CIAR: Interval-based Collaborative Decoding for Image Generation Acceleration [47.52310736101766]
  Auto-regressive (AR) models have recently made notable progress in image generation, achieving performance comparable to diffusion-based approaches.<n>We address this via a cloud-device collaboration framework textbfCIAR, which utilizes on-device self-verification to handle two key properties of visual synthesis.<n>Our solution centers on an on-device token uncertainty quantifier, which adopts continuous probability intervals to accelerate processing and make it feasible for large visual vocabularies.
  arXiv  Detail & Related papers  (2026-03-26T14:04:50Z)
• VAE-REPA: Variational Autoencoder Representation Alignment for Efficient Diffusion Training [53.09658039757408]
  This paper proposes textbfnamex, a lightweight intrinsic guidance framework for efficient diffusion training.<n>name aligns the intermediate latent features of diffusion transformers with VAE features via a lightweight projection layer, supervised by a feature alignment loss.<n>Experiments demonstrate that name improves both generation quality and training convergence speed compared to vanilla diffusion transformers.
  arXiv  Detail & Related papers  (2026-01-25T13:22:38Z)
• Exploring Fourier Prior and Event Collaboration for Low-Light Image Enhancement [1.8724535169356553]
  Event cameras provide performance gain for low-light image enhancement.<n>Currently, existing event-based methods feed a frame and events directly into a single model.<n>We propose a visibility restoration network with amplitude-phase entanglement.<n>In the second stage, a fusion strategy with dynamic alignment is proposed to mitigate the spatial mismatch.
  arXiv  Detail & Related papers  (2025-08-01T04:25:00Z)
• The Devil is in the Darkness: Diffusion-Based Nighttime Dehazing Anchored in Brightness Perception [58.895000127068194]
  We introduce the Diffusion-Based Nighttime Dehazing framework, which excels in both data synthesis and lighting reconstruction.<n>We propose a restoration model that integrates a pre-trained diffusion model guided by a brightness perception network.<n>Experiments validate our dataset's utility and the model's superior performance in joint haze removal and brightness mapping.
  arXiv  Detail & Related papers  (2025-06-03T03:21:13Z)
• FUSE: Label-Free Image-Event Joint Monocular Depth Estimation via Frequency-Decoupled Alignment and Degradation-Robust Fusion [92.4205087439928]
  Image-event joint depth estimation methods leverage complementary modalities for robust perception, yet face challenges in generalizability.<n>We propose the Self-supervised Transfer (PST) and the FrequencyDe-coupled Fusion module (FreDF)<n>PST establishes cross-modal knowledge transfer through latent space alignment with image foundation models, effectively mitigating data scarcity.<n>FreDF explicitly decouples high-frequency edge features from low-frequency structural components, resolving modality-specific frequency mismatches.<n>This combined approach enables FUSE to construct a universal image-event that only requires lightweight decoder adaptation for target datasets.
  arXiv  Detail & Related papers  (2025-03-25T15:04:53Z)
• WTCL-Dehaze: Rethinking Real-world Image Dehazing via Wavelet Transform and Contrastive Learning [17.129068060454255]
  Single image dehazing is essential for applications such as autonomous driving and surveillance. We propose an enhanced semi-supervised dehazing network that integrates Contrastive Loss and Discrete Wavelet Transform. Our proposed algorithm achieves superior performance and improved robustness compared to state-of-the-art single image dehazing methods.
  arXiv  Detail & Related papers  (2024-10-07T05:36:11Z)
• Unrolled Decomposed Unpaired Learning for Controllable Low-Light Video Enhancement [48.76608212565327]
  This paper makes endeavors in the direction of learning for low-light video enhancement without using paired ground truth. Compared to low-light image enhancement, enhancing low-light videos is more difficult due to the intertwined effects of noise, exposure, and contrast in the spatial domain, jointly with the need for temporal coherence. We propose the Unrolled Decomposed Unpaired Network (UDU-Net) for enhancing low-light videos by unrolling the optimization functions into a deep network to decompose the signal into spatial and temporal-related factors, which are updated iteratively.
  arXiv  Detail & Related papers  (2024-08-22T11:45:11Z)
• Enhancing Low-light Light Field Images with A Deep Compensation Unfolding Network [52.77569396659629]
  This paper presents the deep compensation network unfolding (DCUNet) for restoring light field (LF) images captured under low-light conditions. The framework uses the intermediate enhanced result to estimate the illumination map, which is then employed in the unfolding process to produce a new enhanced result. To properly leverage the unique characteristics of LF images, this paper proposes a pseudo-explicit feature interaction module.
  arXiv  Detail & Related papers  (2023-08-10T07:53:06Z)
• FCL-GAN: A Lightweight and Real-Time Baseline for Unsupervised Blind Image Deblurring [72.43250555622254]
  We propose a lightweight and real-time unsupervised BID baseline, termed Frequency-domain Contrastive Loss Constrained Lightweight CycleGAN. FCL-GAN has attractive properties, i.e., no image domain limitation, no image resolution limitation, 25x lighter than SOTA, and 5x faster than SOTA. Experiments on several image datasets demonstrate the effectiveness of FCL-GAN in terms of performance, model size and reference time.
  arXiv  Detail & Related papers  (2022-04-16T15:08:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.