AR Suturing Skill Training Sandbox

ISEF Category: Biomedical Engineering

Subcategory: Other  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: Full Year

The Hook

A surgeon’s hands can look calm while a tiny stitch is off by a few millimeters. That gap can change how well tissue closes. You can turn that idea into a science fair project with an AR training sandbox. Your job is to see whether real-time feedback helps beginners suture more consistently.

What Is It?

This project asks whether augmented reality can help students learn suturing better. Augmented reality, or AR, adds digital information on top of the real world. In your setup, a webcam watches a fake skin surface, like a silicone glove or training pad, while software checks where each stitch lands and how far apart the stitches are.

Think of it like a practice driving app for suturing. The needle and thread are real, but the screen gives live hints, like a lane guide or score meter. Computer vision, which means software that reads images, can measure stitch spacing, angle, and placement. Machine learning, or ML, can then turn those measurements into a skill score.

Why This Is a Good Topic

This is a strong science fair topic because you can test one clear question, whether feedback improves performance. You can measure stitch spacing, consistency, and error rate, so your results are numeric instead of vague. The project connects to real surgical training, where beginners need safe practice before they touch patients. You can also adjust the feedback style, the tracking method, or the scoring system, which gives you room for original work.

Research Questions

• How does real-time visual feedback affect stitch spacing consistency in beginner users?
• What is the effect of AR overlay guidance on the average distance between sutures?
• Does machine-learning-based scoring improve performance more than simple line-and-gap feedback?
• To what extent does webcam tracking accuracy change when the lighting or glove color changes?
• Which feedback style, live color cues, numeric scores, or error highlights, helps users reduce placement mistakes the most?
• How does practice with the AR trainer change performance across repeated trials?

Basic Materials

• Laptop or desktop computer with Unity support.
• Webcam with at least 1080p video.
• Silicone glove, silicone sheet, or commercial suture pad.
• Basic suture kit with needle holder, forceps, scissors, needle, and thread.
• Tape, clips, or a small stand to hold the practice surface steady.
• Printed ruler or calibration target for camera alignment.
• Bright desk lamp or ring light for consistent lighting.
• Notebook or spreadsheet for scoring results.

Advanced Materials

• Portable HMD or phone-based AR device for deployment testing.
• Higher-resolution webcam or depth camera for improved tracking.
• Custom silicone skin models with different textures or thicknesses.
• Force sensor or tension sensor for measuring pull force during suturing.
• Motion capture markers or fiducial markers for pose tracking.
• Medical training suture sets for standardized trials.
• 3D-printed mounting jig for repeatable camera and glove placement.
• Computer with GPU support for training or testing ML models.

Software & Tools

• Unity: Builds the AR interface and handles the user interaction layer.
• WebXR: Runs browser-based augmented reality on supported devices.
• OpenCV: Tracks stitches, markers, and glove landmarks from webcam video.
• ImageJ: Measures spacing and alignment from captured frames when you need a second check.
• Python: Processes data, runs statistics, and compares performance across groups.

Experiment Steps

1. Define the skill you will score first, such as spacing, alignment, tension, or speed.
2. Choose one tracking method, then decide what image features the software must detect reliably.
3. Design a control condition, so you can compare AR feedback against no feedback or delayed feedback.
4. Build a scoring plan that turns image data into measurable performance metrics.
5. Plan how you will test different users, practice rounds, or feedback styles without changing too many variables at once.
6. Decide how you will validate the AR score against a human rater or a separate measurement method.

Common Pitfalls

• Using inconsistent lighting, which makes webcam tracking misread the stitch location from one session to the next.
• Letting the silicone glove shift on the table, which changes the reference frame and ruins spacing measurements.
• Scoring too many skills at once, which makes it hard to tell whether feedback helped stitch placement, spacing, or speed.
• Training the ML model on too few examples, which can make the score look good on test data but fail on new users.
• Comparing users without matching practice time, which can make experience level, not the AR system, drive the results.

What Makes This Competitive

A stronger version of this project does more than build a demo. You compare at least two feedback designs, then show which one improves a specific suturing metric. You also validate your computer-vision score against a human expert or a second measurement method. If you test across different lighting, glove textures, or user skill levels, your work starts to look like real engineering research.

Project Variations

• Test whether the same AR feedback works better on a flat suture pad than on a curved silicone glove.
• Compare beginner users, such as first-time students, with students who already have basic sewing or crafting experience.
• Swap the live visual overlay for post-trial feedback and see whether immediate correction or delayed review improves learning more.

Learn More

• NIH MedlinePlus: Read plain-language background on sutures, wound closure, and medical training topics, then search related articles from the site.
• PubMed: Search review articles on surgical simulation, skill assessment, and computer vision in medical training.
• NIH Library of Medicine 3D Print Exchange: Explore anatomical models and training resources that can inspire your simulator design.
• OpenCV Documentation: Find tutorials for object tracking, calibration, and image measurement in webcam video.
• MIT OpenCourseWare: Look for free courses in computer vision, human-computer interaction, and machine learning that help with the software side.
• Journal of Surgical Education: Search for studies on suturing training, simulation, and performance scoring.