The First Controllable Bokeh Rendering Challenge at NTIRE 2026

2026-05-06

AI Summary

This work addresses the challenge of controllable, high-quality bokeh rendering—particularly the difficulty of simultaneously achieving quantitative fidelity and perceptual realism in portrait and complex foreground scenes. To advance the field, the authors organized the first Controllable Bokeh Rendering Challenge, which attracted 44 participating teams, eight of which submitted valid solutions. The study introduces a dual-track evaluation framework combining objective metrics with expert subjective assessments and conducts comprehensive benchmarking using deep learning–based image synthesis techniques built upon an extended Bokehlicious baseline. The results validate the efficacy of multiple state-of-the-art approaches and establish the first public benchmark for controllable bokeh rendering, thereby promoting standardization and methodological progress in controllable photographic post-processing.

Abstract

This study presents the outcomes of the first Controllable Bokeh Rendering Challenge at NTIRE and highlights the most effective submitted methodologies. In total, 44 participants registered for the competition, of which 8 teams submitted valid solutions after the conclusion of the final test phase. All submissions were evaluated on unseen images, focusing on portraits and intricate subjects with complex and visually appealing bokeh phenomena. In addition to the first track focusing on established quantitative fidelity metrics, we conducted a qualitative user study with a panel of experts for a second track focusing on perceptual assessment. As this was the inaugural challenge on this topic, most of the participants focused on refining and extending the Bokehlicious baseline method.

Problem

• controllable bokeh rendering
• bokeh
• image synthesis
• computational photography

Innovation

• Controllable Bokeh Rendering
• NTIRE Challenge
• Perceptual Assessment
• Bokehlicious Baseline
• Quantitative Fidelity Metrics