Michael E. Bazakos

Physiology-Based Face Recognition in the Thermal Infrared Spectrum

The current dominant approaches to face recognition rely on facial characteristics that are on or over the skin. Some of these characteristics have low permanency can be altered, and their phenomenology varies significantly with environmental factors (e.g., lighting). Many methodologies have been developed to address these problems to various degrees. However, the current framework of face recognition research has a potential weakness due to its very nature. We present a novel framework for face recognition based on physiological information. The motivation behind this effort is to capitalize on the permanency of innate characteristics that are under the skin. To establish feasibility, we propose a specific methodology to capture facial physiological patterns using the bioheat information contained in thermal imagery. First, the algorithm delineates the human face from the background using the Bayesian framework. Then, it localizes the superficial blood vessel network using image morphology. The extracted vascular network produces contour shapes that are characteristic to each individual. The branching points of the skeletonized vascular network are referred to as thermal minutia points (TMPs) and constitute the feature database. To render the method robust to facial pose variations, we collect for each subject to be stored in the database five different pose images (center, midleft profile, left profile, midright profile, and right profile). During the classification stage, the algorithm first estimates the pose of the test image. Then, it matches the local and global TMP structures extracted from the test image with those of the corresponding pose images in the database. We have conducted experiments on a multipose database of thermal facial images collected in our laboratory, as well as on the time-gap database of the University of Notre Dame. The good experimental results show that the proposed methodology has merit, especially with respect to the problem of low permanence over time. More importantly, the results demonstrate the feasibility of the physiological framework in face recognition and open the way for further methodological and experimental research in the area

Automatic detection of vehicle occupants: the imaging problem and its solution

Abstract. The automatic detection and counting of vehicle occupants is a challenging research problem that was given little attention until recently. An automated vehicle-occupant-counting system would greatly facilitate the operation of freeway lanes reserved for car pools (high occupancy vehicle lanes or HOV lanes). There are three major aspects of this problem: (a) the imaging aspect (sensor phenomenology), (b) the pattern recognition aspect, and (c) the system architecture aspect. In this paper, we present a solution to the imaging aspect of the problem. We propose a novel system based on fusion of near-infrared imaging signals and we demonstrate its adequacy with theoretical and experimental arguments. We also compare our solution to other possible solutions across the electromagnetic spectrum, particularly in the thermal infrared and visible regions.

Dictionary learning for robust background modeling

Background subtraction is a fundamental task in many computer vision applications, such as robotics and automated surveillance systems. The performance of high-level visions tasks such as object detection and tracking is dependent on effective foreground detection techniques. In this paper, we propose a novel background modeling algorithm that represents the background as a linear combination of dictionary atoms and the foreground as a sparse error, when one uses the respective set of dictionary atoms as basis elements to linearly approximate/reconstruct a new image. The dictionary atoms represent variations of the background model, and are learned from the training frames. The sparse foreground estimation during the training and performance phases is formulated as a Lasso [1] problem, while the dictionary update step in the training phase is motivated from the K-SVD algorithm [2]. Our proposed method works well in the presence of foreground in the training frames, and also gives the foreground masks for the training frames as a by-product of the batch training phase. Experimental validation is provided on standard datasets with ground truth information, and the receiver operating characteristic (ROC) curves are shown.

Fast access control technology solutions (FACTS)

Positive identification (ID) verification of a large number of people (possibly thousands) in a short time, such as when people enter the gates of a manufacturing plant in the morning, can be a daunting task for the security guards. Current methods that use a password, card swipe system, proximity card reader system, or combination are slow and vulnerable. For example a card can be stolen or a password forgotten, copied or given to an unauthorized person. Hence the need for biometric-based access control systems (ACS). Many biometric systems address the ID verification problem; however, none provide quick and convenient positive (reliable) ID. Fingerprinting is unreliable and requires too much cooperation, face recognition systems (FRS) are unacceptable in a one-to-many application, and iris scan is slow and cumbersome. In this paper, we describe a fast access control technology solution (FACTS) system we developed and demonstrated at Honeywell Labs that enables many motorists to enter a secured parking lot without coming to a complete stop. The FACTS solution combines radio frequency identification (RFID) tags with FRS, reducing the one-to-many FRS problem by using a one-to-one match for facial verification (FV). This solution is dynamic, secure, positive, and hands-free for fast gate access control. At the heart of the Honeywell FV system is a unique tri-band imaging (TBI) camera that reliably locates and records the face of an individual or driver in motion while entering a gate.

Perspective on automatic target recognition evaluation technology

A historical perspective on the evolution of performance evaluation technology in the field of automatic target recognition (ATR) systems is presented. ATR systems have been under development for the last 30 years, and each generation of ATRs and their predecessors have been accompanied by their corresponding stages in the evolution of performance evaluation technology. This perspective will shed light on this field and will answer the questions where are we now and where are we heading?

Activity Awareness: from Predefined Events to New Pattern Discovery

Applying advanced video technology to understand (human) activity and intent, including the interaction of multiple people and objects, is becoming increasingly important, especially for intelligent video surveillance. Recently, technical interest in video surveillance has moved from lowlevel processing modules, such as motion detection and motion tracking, to activity awareness and more complex scene understanding. This paper presents an integrated video surveillance system at Honeywell labs, which detects predefined activities with improved robustness. Also, we present the ‘new activity’ detection (pattern discovery), which can automatically capture new activities, and present the newly detected activities to the operator who checks for their validity and adds them into the activity models. Moreover, we present a torso angle feature, which represents people posture, to detect activities, such as people falling. We used real world data sets to show the effectiveness of our proposed method.

Knowledge- and model-based automatic target recognition algorithm adaptation

One ofthe most critical problems in automatic target recognition (ATR) systems is multiscenario adaptation. Todays ATR systems perform unpredictably, i.e., perform well in certain scenarios and poorly in others. Unless ATR systems can be made adaptable, their utility in battlefield missions remains questionable. We have developed a novel method called knowledge- and model-based algorithm adaptation (KMBAA). KMBAA automatically adapts the ATR parameters as the scenario changes so that ATR can maintain optimum performance. The KMBAA approach has been tested with a nonreal-time ATR simulation system and has demonstrated a significant improvement in detection, false alarm rate reduction, and segmentation accuracy performance.

3D scene modeling using sensor fusion with laser range finder and image sensor

Activity detection (e.g. recognizing peoples behavior and intent), when used over an extended range of applications, suffers from high false detection rates. Also, activity detection limited to 2D image domain (symbolic space) is confined to qualitative activities. Symbolic features, represented by apparent dimensions, i.e. pixels, can vary with distance or viewing angle. One way to enhance performance is to work within the physical space, where object features are represented by their physical dimensions (e.g. inches or centimeters) and are invariant to distance or viewing angle. In this paper, we propose an approach to construct a 3D site model and co-register the video with the site model to obtain real-time physical reference at every pixel in the video. We present a unique approach that creates a 3D site model via fusion of laser range sensor and a single camera. We present experimental results to demonstrate our approach.

Stereopsis And Scene Partitioning For Terrain Interpretation

Determining the planar orientation of the ground surface provides useful additional cues for enhanced performance in various image processing applications. For example, it can provide conflict resolution in target aspect and thus improvement in target classification. It can help in ground vehicle navigation where extremely high slopes or banks would reduce the traversibility of the vehicle. Similarly, it can also be useful in nap-of-the-earth air vehicle navigation through terrain following terrain avoidance. This paper discusses a passive technique to estimate ground plane orientations for scene interpretation. The scene registration problem is handled through a two-dimensional segment-free differential operator. Ranges to points in the scene are computed via an optical flow generation technique and the scene reconstruction is accomplished by a combination of first and second order gradients of the range map.

Automation solutions for the evaluation of plant health in corn fields

The continuously growing need for increasing the production of food and reducing the degradation of water supplies, has led to the development of several precision agriculture systems over the past decade so as to meet the needs of modern societies. The present study describes a methodology for the detection and characterization of Nitrogen (N) deficiencies in corn fields. Current methods of field surveillance are either completed manually or with the assistance of satellite imaging, which offer infrequent and costly information to the farmers about the state of their fields. The proposed methodology promotes the use of small-scale Unmanned Aerial Vehicles (UAVs) and Computer Vision algorithms that operate with information in the visual (RGB) spectrum. Through this implementation, a lower cost solution for identifying N deficiencies is promoted. We provide extensive results on the use of commercial RGB sensors for delivering the essential information to farmers regarding the condition of their field, targeting the reduction of N fertilizers and the increase of the crop performance. Data is first collected by a UAV that hovers over a stressed area and collects high resolution RGB images at a low altitude. A recommendation algorithm identifies potential segments of the images that are candidates exhibiting N deficiency. Based on the feedback from experts in the area a training set is constructed utilizing the initial suggestions of the recommendation algorithm. Supervised learning methods are then used to characterize crop leaves that exhibit signs of N deficiency. The performance of 84.2% strongly supports the potential of this scheme to identify N-deficient leaves even in the case of images where the unhealthy leaves are heavily occluded by other healthy or stressed leaves.