Lesson 5: Avoiding Gait Recognition
Lesson 5: Avoiding Gait Recognition
AI knows how you walk—time to break its pattern.
🔹 TL;DR
🚶 Gait recognition identifies people by their unique walking style.
🤖 Even if your face is hidden, AI can recognize you just from how you move.
🦵 Weighted clothing, misaligned soles, and unpredictable movements can confuse AI.
📢 Change your walk, break the system.
🔍 What’s the Problem?
Even if you cover your face, hair, and outfit, AI can still identify you.
Gait recognition technology analyzes how you walk—your stride length, posture, foot pressure, and movement rhythm. It's already used in:
✔ Law enforcement → Police use it to identify suspects, even if they wear masks.
✔ Smart cities → AI tracks people across multiple locations.
✔ Airports & border control → Some security systems use gait analysis to flag "suspicious" behavior.
👁 AI doesn’t need a clear image of you—blurry security footage is enough for it to match your walk to a database.
📢 Takeaway: AI watches how you move, not just what you look like.
🔗 Example: How Gait Recognition Works
Source: https://link.springer.com/article/10.1007/s11227-020-03409-5
👁 How AI Tracks Your Movement
Gait recognition works by analyzing:
✔ Step length & stride rhythm → The distance and timing between steps.
✔ Foot pressure patterns → How your weight shifts when you walk.
✔ Posture & limb movements → The way your arms swing, your back straightness, and even small habits like limping or toe-walking.
💡 Even small differences in your gait make you identifiable.
📢 Takeaway: If you walk the same way every day, AI will recognize you.
Image source: https://link.springer.com/article/10.1007/s11227-020-03409-5
🛠 What Techniques Can You Use?
To fool gait recognition AI, you need to break the consistency of your walk.
1️⃣ Weighted Clothing – Disrupt Your Natural Movement
🦵 What is it?
Gait recognition AI expects your usual balance and rhythm. Wearing weighted items slightly alters your step timing, making AI less confident in identifying you.
✅ Use:
- Weighted ankle wraps or a weighted vest to subtly shift your step pattern.
- A heavy backpack to force posture changes.
🚫 Avoid:
- Carrying weight the same way every time—AI can adapt if your movement stays consistent.
🔗 Example: Weighted ankle wraps
https://www.sideaita.it/wp-content/uploads/2020/06/0945_Fasce-con-pesi-MY2020.jpg
📢 Takeaway: Changing how your body carries weight can throw AI off track.
2️⃣ Misaligned Soles or Shoe Lifts – Force an Unnatural Gait
👟 What is it?
Gait recognition relies on the symmetry of your steps. Even small changes to your shoes can force AI to misinterpret your walk.
✅ Use:
- One thicker sole or slight heel lift on one foot.
- Inserts or orthotics to change the pressure balance on your feet.
🚫 Avoid:
- Overdoing it—extreme limps can draw unwanted attention from humans, even if they fool AI.
🔗 Example: Shoe Inserts That Change Walking Pressure
📢 Takeaway: A small change to your shoes can cause big errors in AI tracking.
3️⃣ Posture Disruptions – Change the Way You Move
🚶 What is it?
Since gait recognition tracks your natural posture, subtle disruptions can throw AI off.
✅ Use:
- Walk with exaggerated arm swings one day, then hands in pockets the next.
- Hunch slightly or lean to one side while walking.
- Occasionally walk at different speeds or take wider steps.
🚫 Avoid:
- Looking unnatural—you don’t want to draw suspicion from humans.
🔗 Example: How AI Reads Human Posture
Image source: https://blog.tensorflow.org/2019/11/updated-bodypix-2.html
📢 Takeaway: AI expects consistency—break the habit.
4️⃣ Randomized Walking Patterns – Make AI Uncertain
🛤️ What is it?
Gait recognition software gets more confident the longer it watches you. The more random your movement, the less confident AI becomes.
✅ Use:
- Occasionally skip a step, shuffle, or switch pacing.
- Walk at different speeds in different places.
- Turn sharply or change direction suddenly while walking.
🚫 Avoid:
- Making sudden changes only when you see cameras—AI learns to predict this behavior.
📢 Takeaway: The more unpredictable your walking pattern, the harder AI has to work.
⚠ But Does It Work?
✔ AI gait recognition isn’t perfect—it relies on people walking naturally.
✔ Older systems struggle with variations, but newer AI learns patterns over time.
✔ The key is unpredictability—changing your walk randomly keeps AI guessing.
📢 Takeaway: If AI can’t be confident in its match, the data becomes useless.
🛠 How to Use This in Daily Life
✔ Switch up your footwear or add weight to alter your movement.
✔ Use slight posture adjustments to distort tracking.
✔ Walk differently in different locations to create inconsistencies.
✔ Don’t be predictable—variation is your greatest weapon.
📢 Final Takeaway: If AI expects patterns, your job is to break them.
Recap
🚶 Gait recognition AI tracks how you walk, even if your face is hidden.
🦵 Weighted clothing and altered shoes can change your step rhythm.
🚶♂️ Posture shifts and movement changes make AI uncertain.
📢 If AI can’t confidently match your gait, it can’t track you.
🛠 Resources & DIY Guides
🔗 How Gait Recognition Works – ResearchGate
🔗 How AI Reads Human Movement
🔗 Weighted Ankle Wraps for Gait Disruption
What’s Next?
In the next lesson, we’ll look at Masking Your Digital Footprint in Physical Spaces—because AI doesn’t just track your body, it tracks your belongings.
👉 Continue to Lesson 6: Masking Your Digital Footprint in Physical Spaces
👁🗨 Final Thought:
Surveillance AI expects you to walk the same way every day. Change your movement, disrupt the data, and make the system work harder. 🚀