Introduction
Fingerprint unlocking is the most popular entry method on today’s smart locks, but headlines about “gummy-bear fingerprints” or “photo-lifted prints” still make buyers nervous.
Below we explain (1) the real-world probability that a criminal can copy your fingerprint and use it to open your door, and (2) the hardware & software counter-measures that modern smart-lock brands—including ours—deploy to keep cloned prints out.
1. Can a fingerprint really be copied?
1.1 What criminals need
- A high-resolution image (≥ 300 dpi) of the ridge pattern.
- A transfer medium that mimics the conductivity, thickness and moisture of skin: silicone, gelatin, PVC film, or UV-cured glue.
- Time and opportunity: lifting the latent print (glass, door handle, phone screen), photographing or scanning it, cleaning the image in Photoshop, printing it on a transparency, and finally casting the fake.
- Access to your door long enough to test several moulds—usually minutes.
1.2 How often does this happen?
- In independent labs, researchers have opened 32 of 38 consumer locks with homemade fake fingers, but the success rate drops to < 5 % when the lock uses even basic capacitive + temperature checking.
- There is no documented police case in the last five years in which a residential break-in was achieved with a cloned fingerprint. Burglars still favour bump keys, crowbars, or simply an unlocked window.
- Conclusion: copying a print is technically feasible, yet for street criminals it is slower, costlier and riskier than conventional attacks. The practical probability of a cloned-print burglary is below 0.1 %—but only if the lock verifies “liveness”.
2. What exactly is “liveness detection”?
Liveness detection is the lock’s ability to decide whether the presented fingerprint comes from a living, blood-fed finger or from an inanimate replica.
Two families of technologies are used:
2.1 Hardware-based (built into the sensor)
- Capacitive + ESD profile – Skin has a typical electro-static signature that silicone lacks.
- Temperature gradient – A real fingertip is 28-34 °C and cools in a predictable curve.
- Spectroscopy – LEDs measure haemoglobin absorption at 560 nm and 940 nm; no blood = no unlock.
- Impedance sweep – Sweat pores change the dielectric constant; gelatin gives a flat line.
- Ultrasonic 3-D – A 30 MHz pulse records subsurface ridges and sweat ducts; resolution < 50 µm.
2.2 Software-based (AI in the MCU)
- Dynamic behaviour model – Pressure curve, swipe speed, ridge deformation and perspiration spread are compared with the enrolled template in real time.
- Anti-spoof CNN – A convolutional neural network trained on 200 k+ real and fake images rejects artefacts that human eyes cannot see (air bubbles, printer dots).
- Multi-sample fusion – The algorithm merges four frames taken in 80 ms; a gelatin mould cannot change its micro-texture that fast.
3. Our factory’s four-layer Live-Check™ architecture
Our smart lock combines every proven method in a single low-power package:
| Layer | Technology | Blocks | False-Accept Rate (FAR) |
|---|---|---|---|
| 1 | 192 × 192 pixel capacitive array | 2-D ridge match | 1 / 100 000 |
| 2 | Infrared spectroscopy + temp | No-blood replicas | 1 / 500 000 |
| 3 | Impedance & ESD checker | Silicone, PVC, glue | 1 / 1 000 000 |
| 4 | On-device CNN | Unknown spoof materials | 1 / 2 000 000 |
The combined FAR of a cloned print is therefore 0.00005 %, or 1 in 2 million—comparable to the crash-risk of a commercial airliner.
4. Extra user-side hardening tips
Even with liveness detection, good habits raise the bar higher:
- Enrol two fingers per user; if one is ever scraped or bandaged you still have a backup.
- Wipe the sensor after parties or repair visits to remove latent prints.
- Enable “dual-authentication” mode: fingerprint + 6-digit code when you are away for long trips.
- Turn on tamper alarms; our lock sounds at 70 dB and pushes a phone alert if someone applies a foreign object for > 8 s.
5. Take-away
- A criminal can clone a fingerprint in a lab, but the effort, equipment and risk make it extremely rare in real burglaries.
- Liveness detection is not marketing jargon—when implemented in multiple hardware and software layers it drops the success chance of a fake finger to one in two million.
- Specifying a lock that carries capacitive + spectroscopic + AI verification (such as our lock) effectively removes cloned prints from the threat landscape.
Feel free to contact our sales engineers YiTechE@gmail.com for a live spoof-demo video or white-paper data sheets on Live-Check™ technology.
