CLIP-Guided Data Augmentation for Night-Time Image Dehazing

2026-04-07 · Xining Ge, Weijun Yuan, Gengjia Chang, Xuyang Li, Shuhong Liu arxiv

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.

📄 PDF Abstract BibTeX arXiv:2604.05500

Code (0)

등록된 구현이 없습니다.

Tasks

Data Augmentation Image Dehazing

NightHaze: Nighttime Image Dehazing via Self-Prior Learning

2024-03-12 · Beibei Lin, Yeying Jin, Wending Yan, Wei Ye 외

Masked autoencoder (MAE) shows that severe augmentation during training produces robust representations for high-level tasks. This paper brings the MAE-like framework to nighttime image enhancement, demonstrating that se…

Image Dehazing Image Enhancement

Contrastive-SDXL: Annotation-Preserving Night-Time Augmentation for Pedestrian Detection

2026-05-13 · Franky George, Muhammad Khalid, Adil Khan arxiv

Night-time pedestrian detection remains challenging because labelled night-time data are limited and large illumination differences make daytime-only trained detectors unreliable. Latent diffusion models (LDMs) provide a…

Image-to-Image Translation Semantic correspondence Pedestrian Detection

DAP-LED: Learning Degradation-Aware Priors with CLIP for Joint Low-light Enhancement and Deblurring

2024-09-20 · Ling Wang, Chen Wu, Lin Wang

Autonomous vehicles and robots often struggle with reliable visual perception at night due to the low illumination and motion blur caused by the long exposure time of RGB cameras. Existing methods address this challenge …

Autonomous Vehicles Deblurring Depth Estimation Model Optimization

Enhancing Nighttime Vehicle Detection with Day-to-Night Style Transfer and Labeling-Free Augmentation

2024-12-21 · Yunxiang Yang, Hao Zhen, Yongcan Huang, Jidong J. Yang

Existing deep learning-based object detection models perform well under daytime conditions but face significant challenges at night, primarily because they are predominantly trained on daytime images. Additionally, train…

Data Augmentation Generative Adversarial Network object-detection Object Detection+2

Enhancing Vision-Language Models Generalization via Diversity-Driven Novel Feature Synthesis

2024-05-04 · Siyuan Yan, Cheng Luo, Zhen Yu, ZongYuan Ge

Vision-language foundation models like CLIP have shown impressive zero-shot generalization, but finetuning on downstream datasets can cause overfitting and loss of its generalization ability on unseen domains. Although c…

Diversity Zero-shot Generalization