Sequential Gallery for Interactive Visual Design Optimization


†: National Institute of Advanced Industrial Science and Technology (AIST)

‡: The University of Tokyo

Sequential Gallery is an interactive framework for exploring an n-dimensional design space formed by a set of n sliders and then finding an appropriate parameter set from that space. This framework lets the user sequentially select the most preferable option from the options displayed in a grid interface.

To enable this framework, we propose a new Bayesian optimization method called sequential plane search, which decomposes the original high-dimensional search problem into a sequence of two-dimensional search (i.e., plane-search) subtasks.

Video [4:51]

Abstract

Visual design tasks often involve tuning many design parameters. For example, color grading of a photograph involves many parameters, some of which non-expert users might be unfamiliar with. We propose a novel user-in-the-loop optimization method that allows users to efficiently find an appropriate parameter set by exploring such a high-dimensional design space through much easier two-dimensional search subtasks. This method, called sequential plane search, is based on Bayesian optimization to keep necessary queries to users as few as possible. To help users respond to plane-search queries, we also propose using a gallery-based interface that provides options in the two-dimensional subspace arranged in an adaptive grid view. We call this interactive framework Sequential Gallery since users sequentially select the best option from the options provided by the interface. Our experiment with synthetic functions shows that our sequential plane search can find satisfactory solutions in fewer iterations than baselines. We also conducted a preliminary user study, results of which suggest that novices can effectively complete search tasks with Sequential Gallery in a photo-enhancement scenario.

Publication

Yuki Koyama, Issei Sato, and Masataka Goto. 2020. Sequential Gallery for Interactive Visual Design Optimization. ACM Trans. Graph. 39, 4 (July 2020), pp.88:1–88:12. (a.k.a. Proceedings of SIGGRAPH 2020)

DOI: 10.1145/3386569.3392444 (not activated yet)

Download

Related Projects

Sequential Line Search for Efficient Visual Design Optimization by Crowds

Yuki Koyama, Issei Sato, Daisuke Sakamoto, and Takeo Igarashi

ACM Trans. Graph. (SIGGRAPH 2017)

The SIGGRAPH 2020 work proposes a novel method called sequential plane search (SPS) by extending the sequential line search (SLS) method in the SIGGRAPH 2017 work. We experimentally show that our SPS method is much more efficient than the previous SLS method in terms of the number of necessary queries. Both methods are variants of preferential Bayesian optimization (PBO), in which human's relative assessment is involved in the optimization process.

Authors

Yuki Koyama

is a Researcher at National Institute of Advanced Industrial Science and Technology (AIST), Japan. His main research area is Computer Graphics and Human-Computer Interaction, specializing in computational design techniques.

Issei Sato

is an Associate Professor at The University of Tokyo. He works on theory and algorithms in machine learning.

Masataka Goto

is a Prime Senior Researcher at National Institute of Advanced Industrial Science and Technology (AIST), Japan. In 2016, as the Research Director he began a 5-year research project (JST OngaACCEL Project) on music technologies.