Showing papers by "Nalini K. Ratha published in 2016"

Face anti-spoofing with multifeature videolet aggregation

Abstract: Biometric systems can be attacked in several ways and the most common being spoofing the input sensor. Therefore, anti-spoofing is one of the most essential prerequisite against attacks on biometric systems. For face recognition it is even more vulnerable as the image capture is non-contact based. Several anti-spoofing methods have been proposed in the literature for both contact and non-contact based biometric modalities often using video to study the temporal characteristics of a real vs. spoofed biometric signal. This paper presents a novel multi-feature evidence aggregation method for face spoofing detection. The proposed method fuses evidence from features encoding of both texture and motion (liveness) properties in the face and also the surrounding scene regions. The feature extraction algorithms are based on a configuration of local binary pattern and motion estimation using histogram of oriented optical flow. Furthermore, the multi-feature windowed videolet aggregation of these orthogonal features coupled with support vector machine-based classification provides robustness to different attacks. We demonstrate the efficacy of the proposed approach by evaluating on three standard public databases: CASIA-FASD, 3DMAD and MSU-MFSD with equal error rate of 3.14%, 0%, and 0%, respectively.

Visual object recognition

Abstract: Technical solutions are described for training an object-recognition neural network that identifies an object in a computer-readable image. An example method includes assigning a first neural network for determining a visual alignment model of the images for determining a normalized alignment of the object. The method further includes assigning a second neural network for determining a visual representation model of the images for recognizing the object. The method further includes determining the visual alignment model by training the first neural network and determining the visual representation model by training the second neural network independent of the first. The method further includes determining a combined object recognition model by training a combination of the first neural network and the second neural network. The method further includes recognizing the object in the image based on the combined object recognition model by passing the image through each of the neural networks.

Learning face recognition from limited training data using deep neural networks

Abstract: Often deep learning methods are associated with huge amounts of training data. The deeper the network gets, the larger is the need for training data. A large amount of labeled data helps the network learn about the variations it needs to handle in the prediction stage. It is not easy for everyone to get access to huge amounts of labeled data leaving a few to have the luxury to design very deep networks. In this paper, we propose to flatten the disparity by using the modeling methods to minimize the need for huge amounts of data for training a deep network. Using face recognition as an example, we demonstrate how limited labeled data can be leveraged to obtain near state of the art performance with generalization capability across multiple databases. In addition, we show that the normalization in the overall network can improve the speed and resource requirement for the prediction/inferencing stage.

Anonymizing biometric data for use in a security system

Abstract: An anonymized biometric representation of a target individual is used in a computer based security system. A detailed input biometric signal associated with a target individual is obtained. A weakened biometric representation of the detailed biometric signal is constructed such that the weakened biometric representation is designed to identify a plurality of individuals including the target individual. The target individual is enrolled in a data store associated with the computer based security system wherein the weakened biometric representation is included in a record for the target individual. In another aspect of the invention, a detailed input biometric signal from a screening candidate individual is obtained. The detailed biometric signal of the screening candidate is matched against the weakened biometric representation included in the record for the target individual.

Back to the future: A fully automatic method for robust age progression

Abstract: It has been shown that significant age difference between a probe and gallery face image can decrease the matching accuracy. If the face images can be normalized in age, there can be a huge impact on the face verification accuracy and thus many novel applications such as matching driver's license, passport and visa images with the real person's images can be effectively implemented. Face progression can address this issue by generating a face image for a specific age. Many researchers have attempted to address this problem focusing on predicting older faces from a younger face. In this paper, we propose a novel method for robust and automatic face progression in totally unconstrained conditions. Our method takes into account that faces belonging to the same age-groups share age patterns such as wrinkles while faces across different age-groups share some common patterns such as expressions and skin colors. Given training images of K different age-groups the proposed method learns to recover K low-rank age and one low-rank common components. These extracted components from the learning phase are used to progress an input face to younger as well as older ages in bidirectional fashion. Using standard datasets, we demonstrate that the proposed progression method outperforms state-of-the-art age progression methods and also improves matching accuracy in a face verification protocol that includes age progression.

Improving classifier fusion via Pool Adjacent Violators normalization

Abstract: Classifier fusion is a well-studied problem in which decisions from multiple classifiers are combined at the score, rank, or decision level to obtain better results than a single classifier. Subsequently, various techniques for combining classifiers at each of these levels have been proposed in the literature. Many popular methods entail scaling and normalizing the scores obtained by each classifier to a common numerical range before combining the normalized scores using the sum rule or another classifier. In this research, we explore an alternative method to combine classifiers at the score level. The Pool Adjacent Violators (PAV) algorithm has traditionally been utilized to convert classifier match scores to confidence values that model posterior probabilities for test data. The PAV algorithm and other score normalization techniques have studied the same problem without being aware of each other. In this first ever study to combine the two, we propose the PAV algorithm for classifier fusion on publicly available NIST multi-modal biometrics score dataset. We observe that it provides several advantages over existing techniques and find that the interpretation learned by the PAV algorithm is more robust than the scaling learned by other popular normalization algorithms such as min-max. Moreover, the PAV algorithm enables the combined score to be interpreted as confidence and is able to further improve the results obtained by other approaches. We also observe that utilizing traditional normalization techniques first for individual classifiers and then normalizing the fused score using PAV offers a performance boost compared to only using the PAV algorithm.

Outlier faces detector via efficient cohesive subgraph identification

Abstract: A personal or enterprise collection of a large set of face images may contain many types of tags used for querying the collection. Often the tags have many irrelevant content that may not reflect the image content in terms of the facial characteristics. In this paper, we propose a data curation method to filter out the irrelevant face images using a face recognition based subgraph identification. Results on retrievals from the Internet using popular celebrities show the efficacy of our approach after we cleanse the images collection retrieved and applying our algorithm to the collection.

Health monitoring using parallel cognitive processing

Abstract: A system for monitoring participants in a group includes one or more thermal image capturing devices configured to capture one or more thermal images of a plurality of participants in an event, and a processing device configured to receive the one or more thermal images and identification data for at least one of the plurality of participants. The processing device is configured to perform a method that includes identifying the at least one of the plurality of participants, calculating a heat profile of the at least one of the plurality of participants, comparing the heat profile to a reference profile, and determining whether a deviation exists between the heat profile and the reference profile. The method also includes, based on detecting the deviation, calculating a magnitude of the deviation and determining whether a health risk exists based on the magnitude of the deviation.

Securing biometric data through template distribution

Abstract: Embodiments include method, systems and computer program products for safeguarding biometric data. Aspects include receiving a first biometric data unit and generating a template based upon the first biometric data unit. Aspects also include sending the template to a plurality of external template storage devices, each template storage device having a unique device identifier. Aspects also include generating a biometric query including a second biometric data unit. Aspects also include sending the biometric query to at least some of the plurality of external template storage devices. Aspects also include receiving a match score from at least one of the plurality of template storage devices external to the processor, wherein the match score reflects the degree of similarity between the first biometric data unit and the second biometric data unit.

Surface reflectance reduction in images

Abstract: Embodiments include method, systems and computer program products for reducing surface reflectance, for example in a photograph. Aspects include receiving an image set containing a plurality of images of an object. Aspects also include determining geometric transformations for the images and constructing a panorama of the object from the images using the geometric transformations. Aspects also include replacing a portion in the panorama with a corresponding replacement portion from the image set.

Object detection approach using an ensemble strong classifier

Abstract: An object detection method can begin with receipt of an ensemble classifier previously trained upon a binary dataset of images for an object The ensemble classifier can be defined by a list, W, of N weak classifiers, wn for n=1, , N The ensemble classifier can be trained as a detector for the object using a training dataset of images; each image can have a set of candidates having one or more known ground truth candidates As a result of training, the weak classifiers can be reordered, based upon performance, into a new list, W′, and a rejection threshold on candidate rank can be assigned for each detection stage A cascading algorithm can be performed using the trained detector to detect the object in an image having known object candidates The rejection threshold for a stage of the cascading algorithm can be applied to a rank of the object candidates

Proximity feedback for medicine identification

Abstract: Embodiments include method, systems and computer program products for proximity feedback for medication identification. Aspects include accessing, via a user device, a profile data for the user; reading, via the user device, identification data from an identification tag on an object; obtaining information associated with the identification data, wherein the information is associated with contents of the object; and providing a feedback to the user via the user device, the feedback is based upon the information and an interaction between the user of the user device and the object.

System, method, and recording medium for identity fraud prevention in secure transactions using multi-factor verification

Abstract: A secure transaction method, system, and non-transitory computer readable medium, for authorizing a transaction between a user having a personal communication device, a service provider, and a payment provider, include requesting a distribution of a location challenge code to the service provider and a distribution of a biometric data request to the personal communication device of the user, verifying the biometric data of the user based on a match of received biometric data from the user and with biometric data of the user stored in a storage unit, verifying that the location challenge code sent from the personal communication device of the user matches the distributed location challenge code, and sending a verification of authentication of the location challenge code and the match to the payment provider.

Cognitive system to improve athletic performance with motivation from different training styles

Abstract: Techniques for motivating a user during a workout using different coaching styles are provided. In one aspect, a method for motivational coaching of a user during workout sessions includes the steps of: selecting a coaching style for the user based on input from the user and from coaching styles used for at least one other user; determining, during a workout session, whether the coaching style should be changed to enhance performance of the user based on data obtained from the user via a mobile device worn by the user; changing the coaching style if it is determined that the coaching style should be changed to enhance performance of the user; continuing with a current coaching style if it is determined that the coaching style should not be changed; and providing feedback to the user during the workout session based on the coaching style.

Surgical skin lesion removal

Abstract: Embodiments include methods, systems, and computer program products for treating skin lesions. Aspects include receiving an indication that a patient is oriented. Aspects also include acquiring data concerning an area of the patient, the area including a skin lesion. Aspects also include analyzing the data to distinguish between an affected region and an unaffected region of the area of the patient. Aspects also include excising the affected area.

Surface reflectance reduction in images using non-specular portion replacement

Abstract: Embodiments include method, systems and computer program products for reducing surface reflectance, for example in a photograph Aspects include receiving an image set containing a plurality of images of an object Aspects also include determining geometric transformations for the images and constructing a panorama of the object from the images using the geometric transformations Aspects also include replacing a portion in the panorama with a corresponding replacement portion from the image set