The number of applications for cameras in security and monitoring is rising constantly. This trend is supported by improved underlying conditions, such as better resolutions with shrinking image sensors in cameras, increased transfer bandwidths for data, lower energy consumption and increased battery capacities for mobile camera systems like bodycams.

These systems should be reliable. To maintain reliability one also has to trust the underlying data. Even presumed non-existing image data contains information for the evaluator and therefore could be a target for manipulation attempts. But with the growing acceptance and broad distribution of such systems, the effort to prove the falsehood of recorded data is rising.

The reliability and security can be achieved not only with common methods, such as securing the data from outside access (secured servers) or allowing outside access, but also by storing the data in ways that make manipulation attempts costly and most of the time unbearable (distribution of information).

The objective of Trusted Cam is to establish a new and secure means, based on cryptography, of storing image data publicly. The foundation consists of blockchain technologies which aim to secure the reliability and integrity of data on public servers. In comparison to previous technologies, where camera images were saved multiple times in different locations to ensure protection against forgery, Trusted Cam aims to develop protection procedures and features applied directly after recording camera images without storage redundancy.

Commonly used protection features are generally easy-to-generate hash functions, which can rarely be forced and reversed due to the concentration of information. The compactness of the hash and therefore the lower energy and bandwidth needed for subsequent use in comparison to the original image is a clear advantage of hash functions. The low consumption enables the usage of such methods in mobile and wireless systems, which are often highly sensitive to energy and weight.

Furthermore, linking hash values between individual cameras and other system entities enables h mutual protection against the forgery of single cameras.

A Raspberry Pi prototype has been developed in the TrustedCam project. The device can be used to simulate a frame chain for counterfeit-proof surveillance cameras. The frame chain is based on the principles of blockchain technology and distributed systems in general.
For this purpose, the metadata (e.g. hash value, timestamp) of a frame (single image) are embedded in the subsequent frame of the stream. We use the embedded metadata to prove the authenticity of the previous frame. Authenticity is provided by digital signatures that cannot be manipulated. The signatures are generated from the metadata of selected frames and embedded in the subsequent frame.

A robust hash value is determined as part of the metadata to preserve the authenticity of the framechain in case of transmission errors or data compression. Thus, a frame can be identified by the hash value.
In contrast to the cryptographic hash value, the robust hash value is constant with minor changes to the frame and only changes when the semantics is consciously manipulated.

In the future, the computationally intensive processes, such as the embedding of the metadata and the formation of the hash value, will be outsourced to a Field Programmable Gate Array (FPGA). This will allow parallel processing and, therefore, significantly accelerated calculation. The embedding of metadata will also be improved. We will also focus on the readability of embedded metadata in the event of image transmission errors. In addition, we must ensure that the manipulation of the embedded metadata causes a change in the robust hash value.