Iconix: Controlling Semantics and Style in Progressive Icon Grids Generation

Zhida Sun1, Xiaodong Wang1, Zhenyao Zhang1, Min Lu1, Dani Lischinski2, Daniel Cohen-Or3, Hui Huang1*.
1Shenzhen University, 2The Hebrew University of Jerusalem, 3Tel-Aviv University.

Paper arXiv Code Media Coverage

ijhcs_20_teaser

Iconix generates icon grids from a single input concept. The system automates the transition from abstract concepts to a coherent icon continuum by decomposing semantic relationships and performing progressive visual simplification. The result is a structured grid of icons tailored to diverse application scenarios.

Abstract

Visual communication often needs stylistically consistent icons that span concrete and abstract meanings, for use in diverse contexts. We present Iconix, a human-AI co-creative system that organizes icon generation along two axes: semantic richness (what is depicted) and visual complexity (how much detail). Given a user-specified concept, Iconix constructs a semantic scaffold of related analytical perspectives and employs chained, image-conditioned generation to produce a coherent style of exemplars. Each exemplar is then automatically distilled into a progressive sequence, from detailed and elaborate to abstract and simple. The resulting two-dimensional grid exposes a navigable space, helping designers reason jointly about figurative content and visual abstraction. A within-subjects study (N = 32) found that compared to a baseline workflow, participants produced icon grids more creatively, reported lower workload, and explored a coherent range of design variations. We discuss implications for human-machine co-creative approaches that couple semantic scaffolding with progressive simplification to support visual abstraction.

在多样化的设计应用情境中，视觉传达通常需要使用风格统一、且能涵盖具象与抽象含义的图标。本文提出了一种名为 Iconix 的人机共创系统，该系统通过两个维度来组织图标的生成过程：语义丰富度（即描绘的具体内容）和视觉复杂度（即包含的细节多少）。在用户指定特定概念后，Iconix 会构建包含相关分析视角的语义框架，并采用链式、基于图像条件的生成技术，生成一组风格连贯的图标范例。随后，每个范例会被自动提炼为一个渐进序列，呈现出从繁复精细到抽象简约的过渡。由此生成的二维网格暴露出了一个具有极高探索性的设计空间，辅助设计师对具象内容与视觉抽象进行综合考量。一项被试内研究（N=32）表明，与基线工作流相比，参与者能够创作出更具创造性的图标网格，报告了更低的任务负荷，并探索出了更丰富且风格连贯的设计变体。最后，我们探讨了将语义框架与渐进式简化相结合的人机共创方法对支持视觉抽象设计所带来的启示。

The Iconix pipeline

ijhcs_20_detail

Overview of the Iconix pipeline. The pipeline consists of four stages: (a) Concept Ideation identifies related concepts from the input; (b) Semantic Exploration analyzes relations and constructs three-level prompts to generate image exemplars; (c) Visual Simplification transforms the exemplars into icon mask drafts of varying complexity; and (d) Style Refinement optimizes the drafts to generate multi-style icon grids.

The Iconix system

The Iconix system integrates a multi-stage pipeline to support designers in creating progressive and style-consistent icon grids. The user interface is structured into four main modules that guide the designer from initial concept exploration to the final selection of an icon grid.

ijhcs_20_detail

The Iconix’s interface showcases an example of generated results for “fast food”. The system comprises five coordinated modules: (A) Input Concept; (B) Related Concept Analysis (showing candidate ratings); (C) Semantic Exploration (visualizing associations and prompts); (D) Visual Simplification; and (E) the Dual-Axis Icon Grid for comparing design trade-offs.

Evaluation Methods

ijhcs_20_detail

The user study procedure. The study followed a within-subjects design where participants completed two design tasks using both Iconix and the Baseline, with the order of conditions and tasks counterbalanced to ensure fairness.

Results

chi_26_detail

Iconix generates style-consistent icon grids of varying complexity from user-specified concepts. The figure contains 12 concepts: the six on the left are concrete, and the six on the right are abstract. For each concept, icons are organized along two orthogonal dimensions: from top to bottom, semantic richness decreases (from elaborate to simple); from left to right, visual complexity decreases (from detailed to abstract).

chi_26_detail

A gallery of additional icon sets generated by Iconix. The examples demonstrate the system’s ability to produce diverse and coherent outcomes for a range of concrete (left) and abstract (right) concepts, which were selected from prior research and common use.

Reflection

In summary, this research proposes a novel progressive grid-based visual generation approach, leveraging semantic scaffolding to identify relevant analytical perspectives and subsequently synthesize them based on semantic richness and visual complexity. We present Iconix, an AI-assisted system designed to enhance creative ideation in icon design through structured semantic-visual continuums. Through a comprehensive evaluation, we demonstrate how semantic scaffolding and progressive simplification can facilitate abstract concept design in AIGC, enhancing creative workflows and visual ideation.

Video Presentation

coming soon...

BibTeX


                            coming soon...