Lenticular AI Art and Flicker Measurement

ISEF Category: Technology Enhances the Arts

Subcategory: Display Technology  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

A picture can look like it moves, even when the print never changes. That makes lenticular art a perfect test bed for image quality. If you add AI-generated frames, you get a fresh question, how do frame count and viewing angle change flicker? You can turn that question into a real experiment, not just a cool poster.

What Is It?

A lenticular print uses tiny ridges, like a row of narrow window blinds, to send different images to your eyes as you move. When the print contains a sequence of frames, the image seems to animate. The trick is that the lens only shows part of each frame from each angle, so small alignment errors can cause flicker, jumps, or ghosting.

In this project, a diffusion model creates several temporally coherent frames from one still-life prompt. Temporally coherent means the frames should change smoothly, not randomly. You then interlace those frames into one print and view it through a low-cost lenticular sheet. Your job is to test how the number of frames and the viewing angle affect how stable or shaky the animation looks.

Why This Is a Good Topic

This is a strong science fair topic because you can measure it, compare it, and improve it. You are not just making art, you are testing display quality, perception, and image processing. The project connects to real problems in printing, display design, and human visual experience. You can learn prompt design, frame analysis, calibration, and statistics without needing a wet lab.

Research Questions

• How does frame count affect measured flicker in lenticular prints of the same still-life prompt?
• What is the effect of viewing angle on ghosting and frame separation in lenticular AI art?
• Does temporal consistency in the generated frames reduce perceived flicker compared with a less constrained frame set?
• To what extent does interlacing method change image stability across viewing angles?
• Which prompt features, such as object symmetry or background complexity, produce the fewest visible artifacts?
• How does lens pitch or print resolution affect the clarity of frame switching?

Basic Materials

• Computer with internet access for image generation and analysis.
• Access to a diffusion model or image generation tool that can produce multiple related frames.
• Lenticular lens sheet, preferably low-cost and matched to your print resolution.
• Color printer with consistent output settings.
• Photo editing software for frame alignment and interlacing.
• Ruler or angle guide for setting viewing positions.
• Tripod or phone stand for repeatable photography.
• Smartphone camera with manual exposure control.
• Plain paper or poster stock for test prints.
• Tape, scissors, and a cutting mat.

Advanced Materials

• University or maker-space access to a calibrated large-format printer.
• Multiple lenticular lens sheets with different pitches.
• Spectrophotometer or calibrated colorimeter for print consistency checks.
• Motorized rotation stage or angle table for repeatable viewing-angle tests.
• High-resolution camera with fixed focal length lens.
• Controlled lighting setup with diffuse, stable illumination.
• Image processing workstation with Python and OpenCV.
• Printing profile software for color-managed output.
• Flatbed scanner for checking print alignment.
• Computer vision tools for frame registration and artifact detection.

Software & Tools

• Python: Handles image processing, angle-based comparisons, and statistical analysis.
• OpenCV: Measures alignment, contrast changes, and frame-to-frame differences in test images.
• ImageJ: Quantifies visual artifacts and intensity changes across the print.
• Blender: Helps if you want to create or inspect frame sequences and motion paths.
• Google Sheets: Organizes trial data and makes quick plots for early testing.

Experiment Steps

1. Define the visual problem you want to measure, such as flicker, ghosting, or frame switching error.
2. Choose one control factor to vary first, then hold the other print settings fixed.
3. Plan how you will generate frame sets with different levels of temporal coherence.
4. Design a repeatable way to compare viewing angles, print versions, and lens pairings.
5. Build a scoring method or image metric that turns visual artifacts into numbers.
6. Set up a analysis plan that compares conditions with the same prompt and the same lighting.

Common Pitfalls

• Using different lighting for each photo, which changes contrast and makes flicker look worse or better than it really is.
• Skipping lens-to-print alignment checks, which creates false ghosting that comes from offset, not the frame count.
• Comparing prompts with different complexity, which mixes content effects with display effects.
• Letting the diffusion model change object positions too much between frames, which creates motion jumps that look like display failure.
• Testing too few viewing angles, which hides the angle where the lenticular print starts to break down.

What Makes This Competitive

A stronger version of this project goes past a simple before-and-after comparison. You would define a clean metric for flicker or ghosting, then test it across multiple prompts, frame counts, and lens settings. You could also compare human ratings with image-based measures, which adds depth and helps validate your results. Careful controls, clear stats, and a real design rule for making better lenticular art would make the project much stronger.

Project Variations

• Test still-life prompts with different levels of background detail to see which scenes stay clearest through the lens.
• Compare diffusion-generated frames with hand-edited frame sequences to see whether temporal consistency changes artifact levels.
• Swap in different lenticular lens pitches or print resolutions to find the best match for smooth animation.

Learn More

• MIT OpenCourseWare, Computational Photography: search the MIT OpenCourseWare site for courses on image formation, display pipelines, and visual artifacts.
• NASA Image Data and Research: search NASA for free image processing resources and examples of multi-frame visual analysis.
• NIH PubMed: search PubMed for review articles on human visual perception, flicker, and motion perception.
• ImageJ Documentation: find the official ImageJ guides for measuring image intensity, alignment, and contrast.
• Computational Photography, by MIT Press: look for the book at libraries or through preview chapters for core ideas on image capture and display.