Image Credit: “Lines and Minds: Visual Abstraction in Art, Psychology, and Computer Graphics” © 2026 Yael Vinker, Mia Tang, Kartik Chandra, Maneesh Agrawala, Judith Fan, Aaron Hertzmann.
A quick sketch, a diagram on a whiteboard, a comic panel, or a simple icon can often communicate an idea more effectively than a detailed image. Visual abstraction allows us to reduce complexity while preserving meaning, helping people perceive, reason, and communicate in ways that more literal representations sometimes cannot.
At SIGGRAPH 2026, “Lines and Minds: Visual Abstraction in Art, Psychology, and Computer Graphics” brings together researchers and practitioners from multiple disciplines to explore abstraction as a cognitive process, a creative practice, and a computational challenge. As AI systems become increasingly capable of generating realistic imagery, understanding how humans simplify, interpret, and communicate visual information has never been more relevant. We spoke with the Technical Workshop organizers about the enduring power of abstraction and what attendees can expect from this cross-disciplinary conversation.
SIGGRAPH: At its core, visual abstraction is about simplifying without losing meaning. What do you think makes abstraction such a powerful way for people to communicate and understand complex ideas?
Abstraction works because it reduces complexity while preserving, or even amplifying, meaning. The classic example is Picasso’s “Le Taureau” series, where he strips a detailed anatomical bull down to a few essential lines and still captures the animal’s essence. That act of removing rather than adding is what makes it powerful: A simplified form directs attention to what matters and lets the viewer do part of the work of completing the meaning.
SIGGRAPH: This workshop brings together perspectives from cognitive science, art theory, and computer graphics. How do these disciplines approach abstraction differently — and where do you see the most meaningful overlap?
Each discipline has studied abstraction largely in parallel. Cognitive science asks why and how people perceive and produce simplified representations, and what makes them effective for thought and communication. Art and design focus on the choices a maker builds up over years of practice: what to keep, what to leave out, and how expertise shapes those decisions. Computer graphics treats it as a computational challenge: how to build systems that interpret simplified input and generate meaningful visual output. The overlap is that all three are really asking the same underlying question from different angles. Insights from cognitive science and visual theory can inform tools that align with how people actually perceive, and conversely, computational systems give us new ways to test theories about human visual thinking. This Technical Workshop exists to build those bridges.
SIGGRAPH: As AI systems become increasingly skilled at generating detailed, realistic imagery, where does abstraction still play a critical role in creative and communicative work?
Visual abstraction can be an effective tool for communicating intent to a system. Sketches, gestures, and diagrams can sit between human and machine, supporting clearer communication. Artists and designers already leverage visual abstractions like sketches to express and align on a shared creative vision throughout the creative process. Similarly, these abstractions are natural ways for a person to communicate intent to AI systems.
SIGGRAPH: Teaching machines to “abstract” introduces a very different kind of challenge. What makes it so difficult to align computational models with how humans naturally perceive and simplify information?
Deciding what to keep and what to leave out draws on perception, prior knowledge, and communicative goals that humans handle intuitively but that resist being written down as rules. A good simplification is not the same as a compressed or downsampled image; it is a choice about essential structure. So, the open problem is what computational approaches can reveal about how abstraction works, and what stays stubbornly difficult for machines even though it comes naturally to us.
SIGGRAPH: SIGGRAPH brings together artists, researchers, and technologists working at the forefront of visual communication. What makes it an ideal environment for exploring a topic like visual abstraction?
SIGGRAPH’s long-standing interest in visual expression gathers artists, researchers, and technologists in one place, which is exactly the interdisciplinary mix abstraction demands. There’s a clear appetite for it too: Last year’s Drawing and Sketching Technical Workshop drew attendance beyond room capacity. This session continues that effort, following the COGGRAPH workshop at CogSci 2024 and the Drawing and Sketching workshop at SIGGRAPH 2025, and it lets us treat abstraction as both a technical challenge and a shared human-machine language in front of the community best positioned to push it forward.
SIGGRAPH: For attendees joining this session, what perspectives or insights do you hope will resonate beyond the technical workshop itself?
We hope attendees leave seeing abstraction not as a niche stylistic effect but as a cognitive mechanism, a communicative tool, and a computational challenge all at once. The lasting value would be new conversations and collaborations across fields that have studied these questions separately, and a sense that the shared principles underlying how we read and make simplified pictures are worth pursuing together. As humans and machines increasingly create side by side, understanding abstraction shapes how we think, communicate, and build the next generation of more intuitive, human-centered tools.
Interested in exploring the intersection of art, perception, and technology? Be sure to attend “Lines and Minds: Visual Abstraction in Art, Psychology, and Computer Graphics” at SIGGRAPH 2026. View the full schedule to discover more Technical Workshops and Courses shaping the future of computer graphics and interactive techniques.
Dr. Yael Vinker is a postdoctoral associate at MIT CSAIL, working with Prof. Antonio Torralba, and an incoming assistant professor at the Weizmann Institute of Science. Her research focuses on generative models for visual communication at the intersection of computer vision, graphics, machine learning, and design.
Dr. Vinker’s work has been recognized with two SIGGRAPH Best Paper Awards, two SIGGRAPH Honorable Mention Awards, and the MIT EECS Rising Stars distinction.
Mia Tang is a first-year CS Ph.D. student at Stanford University, advised by Professor Maneesh Agrawala. Previously, she earned her Bachelor of Computer Science and Arts from Carnegie Mellon University. Her research explores the intersection of computer graphics, vision, and AI, focusing on developing interactive, controllable AI systems that align with natural human processes.
