Sign in to view source links and access this dataset
Description
Released in 2026 with the CVPR paper 'UniSER: A Foundation Model for Unified Soft Effects Removal', this dataset contains approximately 80,000 unique clean images paired with around 2 million synthetic renderings of haze, fog, and smoke. It covers homogeneous, non-homogeneous, indoor, outdoor, daytime, and dense atmospheric conditions for training and benchmarking single-image dehazing models.
Use Cases
Train single-image dehazing models based on the large-scale synthetic haze renderings.
Benchmark dehazing algorithm performance across diverse conditions like indoor and outdoor scenes.
Develop foundation models for unified soft effects removal based on the physically-motivated renderings.
Strengths
Contains ~2 million synthetic haze/fog/smoke renderings paired with ~80,000 clean images.
Covers a wide range of atmospheric conditions including homogeneous, non-homogeneous, indoor, and outdoor scenes.
Released in conjunction with a CVPR 2026 paper, suggesting a research-grade purpose.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
The dataset size is listed as ~2.5 TB, which requires significant storage and download capacity.
Provenance
Source
Author jdzhang0929 on Hugging Face.
Collection Method
Synthetically generated renderings of haze, fog, and smoke applied to clean images.
Freshness
Last updated 2026-05-27 21:19:55.
License is unknown; terms of use must be verified before application.