HIGH SPEED BIOMETRIC PROBE IMAGE RETRIEVAL USING BIG IMAGE DATA ON THE CLOUD Mohd. Ahmed Abdul Mannan PHD scholar, JJT university, Rajasthan, India ermahmeds@gmail.com Abstract Recent development in digital technology leads to a mass storage of images which include citizen, patient and criminal information like fingerprints, iris, face photographs in digital form, For criminals these information stored across the different cities and for patient Medical imaging technology, already having a cloud services for storing the patient information, like AT&T Medical Imaging. In case if an accident occurs and patient is brought to the hospital also patient is unconscious then his old stored record will be very helpful for getting his past medical history and to inform his relatives. In such cases search through biometric technology like finger print on the cloud is require. While for criminal, once any of the criminal information is found from the crime places like finger print or sketch of the criminal by forensic officer then the same is very useful to find out the criminal, Normally these information’s is kept on a PC at the city or state government agencies. Present method of investigation is taking too much time like in India as of now investigating agencies are sending the finger prints lifted from a scene of crime for matching to state finger print bureau. If no match found, then the prints are sent to the Central Finger Print Bauru (CFPB) and finger print bureaus of different states. This is very time consuming process. Cloud computing and Big data analytics provide large scope to solve such problem. The paper proposes a model and discusses the latest technology through which images of Big dataset, across cities, hospitals can be stored on the cloud by considering the security issues. It further proposes method for retrieving the probe image from database created on the cloud. Keywords: Minutiae, Feature matching, Big data, Cloud computing, Biometric probe Image. 1. INTRODUCTION Searching an image like face photograph, forensic sketch or biometric image like a fingerprint, iris etc. on a large data set was hot topic of last decade and many methods and techniques have been actively implemented for feature extraction, feature matching and Minutiae-based matching. For example SIFT[1][2], PCA-SIFT[3], Harris-SIFT[4], SURF[5], FAST[6], AGAST[7], BRIEF[8], ORB[9] and BRISK[10] etc. are the best known feature detectors and descriptors while global and local minutiae-based matching are used for fingerprint matching. Methods for extracting the feature of face photograph, sketch images etc. basically work in three steps detecting the features, extracting the features then finally matching the features while fingerprint matching is also having a feature extraction technique for minutiae based matching. During the feature detection stage, some interest operator is applied on the images to find distinctive key points, which are likely to match well in other images. During the feature description stage, the detected features are described based on the neighbor pixels around it. Basically there are two types of descriptors: vector descriptor or binary descriptor. Vector descriptor is a feature vector with n dimensions, for instance n=128 for SIFT features. It stores more information, but it is difficult to find the nearest match in high dimensional space. Binary vector is n-bit binary String consisting of 0 and 1. It can be processed quite fast with efficient algorithms. During the matching stage, each query feature is matched to the most similar feature in the reference image based on their descriptors. There are different types of image matching techniques based on querying method. The query methods are Keywords based, Graphics based, Image based and Composite. *Mohd. Ahmed Tel: +91 9987335278, Email Address: ermahmeds@gmail.com