ProMist-5K: A Comprehensive Dataset for Digital Emulation of Cinematic Pro-Mist Filter Effects

TL;DR

ProMist-5K addresses the lack of physically grounded digital diffusion models for Pro-Mist cinematography. It introduces a scene-referred diffusion pipeline with six Gaussian blur layers to emulate halo and diffusion under four configurations, producing 20,000 paired samples. The dataset enables robust training and evaluation of image translation models for cinematic diffusion, with results showing paired methods perform best and demonstrate the dataset's utility for both supervised and unsupervised settings. This resource bridges digital flexibility with traditional lens aesthetics, supporting future work in computational cinematography and post-production.

Abstract

Pro-Mist filters are widely used in cinematography for their ability to create soft halation, lower contrast, and produce a distinctive, atmospheric style. These effects are difficult to reproduce digitally due to the complex behavior of light diffusion. We present ProMist-5K, a dataset designed to support cinematic style emulation. It is built using a physically inspired pipeline in a scene-referred linear space and includes 20,000 high-resolution image pairs across four configurations, covering two filter densities (1/2 and 1/8) and two focal lengths (20mm and 50mm). Unlike general style datasets, ProMist-5K focuses on realistic glow and highlight diffusion effects. Multiple blur layers and carefully tuned weighting are used to model the varying intensity and spread of optical diffusion. The dataset provides a consistent and controllable target domain that supports various image translation models and learning paradigms. Experiments show that the dataset works well across different training settings and helps capture both subtle and strong cinematic appearances. ProMist-5K offers a practical and physically grounded resource for film-inspired image transformation, bridging the gap between digital flexibility and traditional lens aesthetics. The dataset is available at https://www.kaggle.com/datasets/yingtielei/promist5k.

Figures (5)

• Figure 1: Comparison of effects under different densities and focal lengths. At a fixed density, increasing the focal length from (a) to (b) intensifies highlight diffusion. At a fixed focal length, reducing the density from (a) to (c) weakens the softening effect. (d) shows that low density combined with longer focal length maintains a subtle yet visible cinematic atmosphere.
• Figure 2: Pro-Mist filter effect (1/2@20mm): Original vs. filtered image with their 3D HSV space models and 1D histograms. Altered HSV distributions illustrate highlight suppression, contrast reduction, and desaturation.
• Figure 3: Pro-Mist Filter Emulation Pipeline.
• Figure 4: Contribution of Multi-Scale Blur Layers to Pro-Mist Emulation. Increasing the number of blur layers improves the fidelity of halation rendering.
• Figure 5: Qualitative comparison of different translation models on two representative Pro-Mist emulation tasks (Original $\rightarrow$ 1/2@20mm and Original $\rightarrow$ 1/8@20mm).