Trusted Cameras on Mobile Devices Based on SRAM Physically Unclonable Functions

Abstract

Nowadays, there is an increasing number of cameras placed on mobile devices connected to the Internet. Since these cameras acquire and process sensitive and vulnerable data in applications such as surveillance or monitoring, security is essential to avoid cyberattacks. However, cameras on mobile devices have constraints in size, computation and power consumption, so that lightweight security techniques should be considered. Camera identification techniques guarantee the origin of the data. Among the camera identification techniques, Physically Unclonable Functions (PUFs) allow generating unique, distinctive and unpredictable identifiers from the hardware of a device. PUFs are also very suitable to obfuscate secret keys (by binding them to the hardware of the device) and generate random sequences (employed as nonces). In this work, we propose a trusted camera based on PUFs and standard cryptographic algorithms. In addition, a protocol is proposed to protect the communication with the trusted camera, which satisfies authentication, confidentiality, integrity and freshness in the data communication. This is particularly interesting to carry out camera control actions and firmware updates. PUFs from Static Random Access Memories (SRAMs) are selected because cameras typically include SRAMs in its hardware. Therefore, additional hardware is not required and security techniques can be implemented at low cost. Experimental results are shown to prove how the proposed solution can be implemented with the SRAM of commercial Bluetooth Low Energy (BLE) chips included in the communication module of the camera. A proof of concept shows that the proposed solution can be implemented in low-cost cameras.

Secure Combination of IoT and Blockchain by Physically Binding IoT Devices to Smart Non-Fungible Tokens Using PUFs

“…This bit flipping is generally unwanted when the SRAM cell start-up value is used in the PUF response, since the PUF reproducibility is reduced and the complexity of the error correction code raises accordingly. To mitigate this, the solution explained in [ 32 ] is used herein, which only employs the most stable cells to generate the SRAM PUF response. If the resulting SRAM PUF responses show some bias, a debiasing algorithm has to be considered.…”

“…The unbiased stable cells after this classification process are named ID cells. Using these ID cells, a simple repetition error correction code works correctly, as shown in [ 16 , 32 ].…”

Efficient Source Camera Identification with Diversity-Enhanced Patch Selection and Deep Residual Prediction

“…Images can be easily edited to cover up information for illegal purposes, and it can be difficult to distinguish edits with the naked eye. Therefore, tools for image forensics are in urgent need to verify the provenance and authenticity of images [ 1 , 2 , 3 ].…”