人工智能指纹识别技术在警务实战中的应用

AI Fingerprint Recognition and Its Policing Practical Utilization

大数据时代,海量指纹数据的汇聚对指纹识别系统的比对精度和速度带来了挑战。能否实现对现场指纹图片在海量样本库中的“以图搜图”,即无需对指纹图片进行特征编辑即可快速比对(无特征比对),是一线部门提出的实战需求。本文以“云痕”智能指纹识别系统为例,介绍一种将人工智能引入指纹识别领域,采用自适应小波框架与主动式深度学习的比对技术和总线级协同计算的计算技术,建设指纹亿级数据库的技术创新,使得指纹的比对精度和速度得到提升。结合笔者单位警务工作,介绍了“云痕”系统应用于电脑端和移动端的指纹比对场景,对比分析了人工智能与传统指纹识别系统在指纹表述信息量及比对精度等方面的技术优势,并对人工智能在指纹识别方面的应用前景进行了展望。

Great challenges have been being brought to the fingerprint identification systems into their matching accuracy and speed with the so-called big-data collection and entries of billions of fingerprints.Forensic experts and practitioners expect an automatic fingerprint recognition technology(also known as non-minutiae-based matching)would be applied so as to eligibly search out the matched fingerprint from the fingerprint gallery with just an intact shot-on-the-scene image of fingerprint.YUNHEN,an intelligent fingerprint identification system,has thereby been smartly created for coping against the above-indicated challenges.It is an innovative facility,taking the advantages of self-adaptive wavelet algorithmic framework,proactive deep learning and BUS synergy,so that it can realize the approvable accuracy and speed of fingerprint matching among a billion-level data gallery of fingerprint.Exampled with the policing practical utilization,YUNHEN system was here introduced into its actual scenarios of fingerprint matching applied throughout the ends of computer and mobile phone.Accordingly,such an artificial intelligent operational facility was compared and analyzed against the traditional fingerprint identification system on terms of delivered fingerprint information capacity and accuracy,demonstrating its overwhelming advantages.Finally,prospect was envisioned about the application of artificial intelligence into fingerprint identification.