Multiple granularity fused mobile forensics algorithm

• Multiple granularity framework is introduced to person forensics system.
• A bionic pattern recognition based face recognition algorithm could achieve a high precision in a fairly short time.
• Two kinds of fusion strategies are compared for the application on mobile forensics systems.

With the rapid development of Internet of Things and mobile computing platforms, many intelligent applications are implemented as mobile services, such as pattern recognition algorithms. This paper proposes two kinds of multiple granularity feature-based person recognition algorithms for forensics applications, which fuse biometric feature and face recognition algorithms into two kinds of strategies. A bionic pattern recognition based face recognition algorithm is also proposed. Experimental results show that the proposed algorithm has a satisfied recognition precision and low time complexity in many forensics scenarios, which could be conducted on mobile forensics devices with high efficiency.