We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
The Living Dead: Hacking Mobile Face Recognition SDKs with Non-Deepfake Attacks
Discover how hackers are exploiting face recognition SDKs with non-deepfake attacks, compromising liveness detection and creating fake accounts for malicious purposes.
- Even in the AI era, face recognition SDKs can still be hacked.
- Liveness detection, supposedly more secure, requires interactions between the app and the cloud.
- 18 face recognition SDKs were analyzed, and 11 were found to have insecure design or implementation.
- Attackers can bypass face recognition software by replacing the result with a fake image or video.
- Some SDKs return unencrypted image data, making it vulnerable to attacks.
- Hooking is used to bypass liveness detection on device and conduct attacks.
- Face recognition SDKs can be compromised to create fake accounts for tax evasion, etc.
- Commercial packers can defeat face recognition software using various techniques.
- Insecure face recognition SDKs are commonly used in digital wallets, etc.
- Attackers can capitalize on flaws in system design or implementation to create fake accounts.
- Face recognition software relies on liveness detection, which can be bypassed using attacks.
- SDKs may sacrifice more security due to poor design or implementation.
- Face recognition SDKs can be compromised to create fake accounts for various purposes.
- Some SDKs conduct environment checking but can still be bypassed using attacks.
- Face recognition software can be compromised using various techniques, including presentation attacks.
- Liveness detection can be bypassed using fake image or video.
- Face recognition SDKs can be compromised to pass the passport check.
- Hackers can create fake accounts using various techniques.
- Face recognition software can be compromised using various attacks, including spoofing and liveness detection bypassing.
- Fake image or video can be used to bypass face recognition software.
- Face recognition SDKs can be compromised to create fake accounts for various purposes.