While the whole world goes digital, law enforcements across the world are still faced with task of processing paper fingerprint cards and conversion it to digital form. It’s not only about digitizing of archives stored for many years, but also about automation of daily work, because in many locations fingerprinting is still ink&roll.
Conversion of fingerprint cards is still a tedious process, consuming a lot of time from an officer. Very often an operator has to do pre-settings, manual adjustments or even insert some information and “help to detect” fingerprints manually. These routine operations waste a lot of time of law enforcements, which could have been spent for real work.
Neurodactyl introduces new SDK combining different machine learning technologies for quick and fully automated digitizing of fingerprint cards. The SDK processes any type of images, regardless the fingerprint card format, including filled in manually on blank paper. The algorithms accurately detect, segment and align flat and rolled fingerprints both for their further passing to an AFIS. The SDK also automatically detects missing fingers, right and left hand and checks, if slap and rolled impressions are belong to the same person and that their positions are correct. AI algorithms provide accurate detection and segmentation, even with artifacts on fingerprints such as crossing lines, hand writings, etc. and with zero manual pre-settings from a user.
The SDK provides very easy interface with few calls for quick integration in an AFIS. The SDK can utilize GPU and GPU both, providing high speed of processing for many images in batch mode on GPU.
High accuracy of detection, segmentation and recognition allows to process correctly vast majority of fingerprint cards without any manual correction, which saves a lot of operator’s time and money of the customers.
Conversion of fingerprint cards is still a tedious process, consuming a lot of time from an officer. Very often an operator has to do pre-settings, manual adjustments or even insert some information and “help to detect” fingerprints manually. These routine operations waste a lot of time of law enforcements, which could have been spent for real work.
Neurodactyl introduces new SDK combining different machine learning technologies for quick and fully automated digitizing of fingerprint cards. The SDK processes any type of images, regardless the fingerprint card format, including filled in manually on blank paper. The algorithms accurately detect, segment and align flat and rolled fingerprints both for their further passing to an AFIS. The SDK also automatically detects missing fingers, right and left hand and checks, if slap and rolled impressions are belong to the same person and that their positions are correct. AI algorithms provide accurate detection and segmentation, even with artifacts on fingerprints such as crossing lines, hand writings, etc. and with zero manual pre-settings from a user.
The SDK provides very easy interface with few calls for quick integration in an AFIS. The SDK can utilize GPU and GPU both, providing high speed of processing for many images in batch mode on GPU.
High accuracy of detection, segmentation and recognition allows to process correctly vast majority of fingerprint cards without any manual correction, which saves a lot of operator’s time and money of the customers.
