This paper addresses the accuracy problem of an eye gaze tracking system. We first analyze the technical barrier for a gaze tracking system to achieve a desired accuracy, and then propose a subpixel tracking method to break this barrier. We present new algorithms for detecting the inner eye corner and the center of an iris at subpixel accuracy, and we apply these new methods in developing a real-time gaze tracking system. Experimental results indicate that the new methods achieve an average accuracy within 1.4/spl deg/ using normal eye image resolutions.

/pdf/subpixel-eye-gaze-tracking-28cjvt6fog.pdf

Subpixel eye gaze tracking

Determination of relative three-dimensional (3–D) position, orientation, and relative motion between two reference frames is an important problem in robotic guidance, manipulation, and assembly as well as in other fields such as photogrammetry. A solution to this problem that uses two-dimensional (2–D), intensity images from a single camera is desirable for real-time applications. Where the object geometry is unknown, the estimation of structure is also required. A single camera is advantageous because a standard video camera is low in cost, setup and calibration are simple, physical space requirements are small, reliability is high, and low-cost hardware is available for digitizing and processing the images. A difficulty in performing this measurement is the process of projecting 3–D object features to 2–D images, a nonlinear transformation. Noise is present in the form of perturbations to the assumed process dynamics, imperfections in system modeling, and errors in the feature locations extracted from the 2–D images. This dissertation presents solutions to the remote measurement problem for a dynamic system given a sequence of 2–D intensity images of an object where feature positions of the object are known relative to a base reference frame and where the feature positions are unknown relative to a base reference frame. The 3–D transformation is modeled as a nonlinear stochastic system with the state estimate providing six degree-of-freedom motion and position values. The stochastic model uses the iterated extended Kalman filter as an estimator and as a screw representation of the 3–D transformation based on dual quaternions. Dual quaternions provide a means to represent both rotation and translation in a unified notation. The method has been implemented and tested with both simulated and actual experimental data. Simulation results are provided along with comparisons to a point-based method using rotation and translation to show the relative advantages of the proposed method. Experimental test results using a camera mounted on the end effector of a robot arm to demonstrate relative motion and position control are also presented. iv

Pose and motion estimation from vision using dual quaternion-based extended kalman filtering

https://escholarship.org/content/qt5pc374hf/qt5pc374hf.pdf?t=omjkkq

Optimizing disk registration algorithms for nanobeam electron diffraction strain mapping

Purpose: To describe a novel computer-based image analysis method that is being developed to assist and automate the diagnosis of retinal disease. Methods: Content-based image retrieval is the process of retrieving related images from large database collections using their pictorial content. The content feature list becomes the index for storage, search, and retrieval of related images from a library based upon specific visual characteristics. Low-level analyses use feature description models and higher-level analyses use perceptual organization and spatial relationships, including clinical metadata, to extract semantic information. Results: We defined, extracted, and tested a large number of region- and lesion-based features from a dataset of 395 retinal images. Using a statistical hold-one-out method, independent queries for each image were submitted to the system and a diagnostic prediction was formulated. The diagnostic sensitivity for all stratified levels of age-related macular degeneration ranged from 75% to 100%. Similarly, the sensitivity of detection and accuracy for proliferative diabetic retinopathy ranged from 75% to 91.7% and for nonproliferative diabetic retinopathy, ranged from 75% to 94.7%. The overall purity of the diagnosis (specificity) for all disease states in the dataset was 91.3%. Conclusions: The probabilistic nature of content-based image retrieval permits us to make statistically relevant predictions regarding the presence, severity, and manifestations of common retinal diseases from digital images in an automated and deterministic manner.

Automated diagnosis of retinopathy by content-based image retrieval.

Array beamforming techniques allow for the generation of 3-D spatial filters which can be used to localize objects in a large field of view (FOV) without the need for mechanical scanning. By combining broadband beamforming with a sparse, random array of receivers, we have constructed a low-cost, yet powerful, in-air sonar system, which is suited for a wide range of robotic applications. Experimental results in unmodified office environments show the performance of the sonar sensor. In particular, we document the sensor's capacity to produce 3-D location measurements in the presence of multiple highly overlapping echoes. We show how this capability makes possible the combination of a wide FOV with accurate 3-D localization, allowing the sensor to operate under real-time constraints in realistic environments. To demonstrate the use of this sensor, we describe an odometry application that estimates egomotion of a mobile robot using acoustic flow.

Broadband 3-D Sonar System Using a Sparse Array for Indoor Navigation

A digiuil image processing inspection system is under development at Oak Ridge National Laboratory that will locateimage features on printed material and measure distances between them to accuracies of 0.001 in. An algorithm has beendeveloped for this system that can locate unique image features to subpixel accuracies. It is based on a least-squares fit ofa paraboloid function to the surface generated by correlating a reference image feature against a test image search area.Normalizing the correlation surface makes the algorithm robust in the presence of illumination variations and local flaws.Subpixel accuracies of better than 1/16 of a pixel have been achieved using a variety of different reference image features. 1. INTRODUCTIONAn algorithm has been developed at Oak Ridge National Laboratory (ORNL) that can be used to locate a variety offeatures or objects in a digitized image to subpixel accuracies. The algorithm uses a set of normalized correlation j1 generated by correlating a stored reference template against a test image search area that is believed to contain the feature

/pdf/subpixel-measurement-of-image-features-based-on-paraboloid-4ng21xfoi5.pdf

Subpixel measurement of image features based on paraboloid surface fit

This paper describes current work at the Oak Ridge National Laboratory to develop a robotic vision system capable of recognizing designated objects by their intrinsic geometry. This method, based on single camera vision, combines point features and a model-based technique using geometric feature matching for the pose calculation. In this approach, 2-D point features are connected into higher-order shapes and then matched with corresponding features of the model. Pose estimates are made using a closed-form point solution based on model features of four coplanar points. Rotations are represented by quaternions that simplify the calculations in determining the least squares solution for the coordinate transformation. This pose determination method including image acquisition, feature extraction, feature correspondence, and pose calculation has been implemented on a real-time system using a standard camera and image processing hardware. Experimental results are given for relative error measurements.

Robust pose determination for autonomous docking

A digital machine-inspection system is being developed at Oak Ridge National Laboratory to detect flaws on printed graphic images. The inspection is based on subtraction of a digitized test image from a reference image to determine the location, number, extent, and contrast of potential flaws. When performing subtractive analysis on the digitized information, two sources of errors in the amplitude of the difference image can develop: (1) spatial misregistration of the reference and test sample, or (2) random fluctuations in the printing process. Variations in printing and registration between samples will generate topological artifacts related to surface structure, which is referred to as edge noise in the difference image. Most feature extraction routines require that the difference image be relatively free of noise to perform properly. A novel algorithm has been developed to filter edge noise from the difference images. The algorithm relies on the a priori assumption that edge noise will be located near locations having a strong intensity gradient in the reference image. The filter is based on the structure of the reference image and is used to attenuate edge features in the difference image. The filtering algorithm, consisting of an image multiplication, a global intensity threshold, and an erosion/dilation, has reduced edge noise by 98% over the unfiltered image and can be implemented using off-the-shelf hardware.© (1990) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Nonlinear filter derived from topological image features

Under a program sponsored by the Department of Energy, the Oak Ridge complex is developed a ``Portal-of-the-Future``, or ``smart portal``. This is a security portal for vehicular traffic which is intended to quickly detect explosives, hidden passengers, etc. It uses several technologies, including microwaves, weigh-in-motion, digital image processing, and electroacoustic wavelet-based heartbeat detection. A novel component of particular interest is the Enclosed Space Detection System (ESDS), which detects the presence of persons hiding in a vehicle. The system operates by detecting the presence of a human ballistocardiographic signature. Each time the heart beats, it generates a small but measurable shock wave that propagates through the body. The wave, whose graph is called a ballistocardiogram, is the mechanical analog of the electrocardiogram, which is routinely used for medical diagnosis. The wave is, in turn, coupled to any surface or object with which the body is in contact. If the body is located in an enclosed space, this will result in a measurable deflection of the surface of the enclosure. Independent testing has shown ESDS to be highly reliable. The technologies used in the smart portal operate in real time and allow vehicles to be checked through the portal in much less time than would be required for human inspection. Although not originally developed for commercial transportation, the smart portal has the potential to solve several transportation problems. It could relieve congestion at international highway border crossings by reducing the time required to inspect each vehicle while increasing the level of security. It can reduce highway congestion at the entrance of secure facilities such as prisons. Also, it could provide security at intermodal transfer points, such as airport parking lots and car ferry terminals.

/pdf/application-of-the-smart-portal-in-transportation-1bnmv59tll.pdf

Application of the smart portal in transportation

Engineers at Oak Ridge National Laboratory have been investigating the feasibility of computer-controlled docking in resupply missions, sponsored by the US Army. The goal of this program is to autonomously dock an articulating robotic boom with a special receiving port. A video camera mounted on the boom provides video images of the docking port to an image processing computer that calculates the position and orientation (pose) of the target relative to the camera. The control system can then move the boom into docking position. This paper describes a method of uniquely identifying and segmenting the receiving port from its background in a sequence of video images. An array of light-emitting diodes was installed to mark the vertices of the port. The markers have a fixed geometric pattern and are modulated at a fixed frequency. An asynchronous demodulation technique to segment flashing markers from an image of the port was developed and tested under laboratory conditions. The technique acquires a sequence of images and digitally processes them in the time domain to suppress all image features except the flashing markers. Pixels that vary at frequencies within the filter bandwidth are passed unattenuated, while variations outside the passband are suppressed. The image coordinates of the segmented markers are computed and then used to calculate the pose of the receiving port. The technique has been robust and reliable in a laboratory demonstration of autodocking.

/pdf/an-asynchronous-modulation-demodulation-technique-for-robust-3cjzyx56t9.pdf

William Bruce Jatko

Papers

Subpixel measurement of image features based on paraboloid surface fit

Robust pose determination for autonomous docking

Nonlinear filter derived from topological image features

Application of the smart portal in transportation

An asynchronous modulation/demodulation technique for robust identification of a target for 3-D pose determination