This paper proposes a method that can spot and recognize each facial expression from time-sequential images that contain multiple facial expressions that could abruptly change from one expression to another expression...
详细信息
ISBN:
(纸本)0818684976
This paper proposes a method that can spot and recognize each facial expression from time-sequential images that contain multiple facial expressions that could abruptly change from one expression to another expression. Previously, the authors have proposed an HMM (Hidden Markov Models) based method Sor recognizing a spotted facial expression. In this paper, to HMM, we add states corresponding to the simultaneous motion. of two different facial expressions: i.e. a muscle relaxation for one expression and a muscle contraction for another expression. Then, the added states are each linked from the HMM apex state of one expression and are linked to that of another expression. Experimental results showed that for most pairs of expressions the change in expression can be recognized accurately. In addition, recognition rate for very fast change of expressions improved significantly. The proposed method was applied to regenerate facial expressions on a synthesized character to show the method's effectiveness in obtaining facial motion information.
This paper addresses large-displacement-diffeomorphic mapping registration from an optimal control perspective. This viewpoint leads to two complementary formulations. One approach requires the explicit computation of...
详细信息
ISBN:
(纸本)9781424439942
This paper addresses large-displacement-diffeomorphic mapping registration from an optimal control perspective. This viewpoint leads to two complementary formulations. One approach requires the explicit computation of coordinate maps, whereas the other is formulated strictly in the image domain (thus making it also applicable to manifolds which require multiple coordinate charts). We discuss their intrinsic relation as well as the advantages and disadvantages of the two approaches. Further we propose a novel formulation for unbiased image registration, which naturally extends to the case of time-series of images. We discuss numerical implementation details and carefully evaluate the properties of the alternative algorithms.
In this paper we discuss and analyze possible futures for technologies in the field of computervision (CV). Using a method we have coined speculative analysis we take a broad look at research trends in the field to c...
详细信息
ISBN:
(纸本)9781538607336
In this paper we discuss and analyze possible futures for technologies in the field of computervision (CV). Using a method we have coined speculative analysis we take a broad look at research trends in the field to categorize risks, analyze which ones are most threatening and likely, and ultimately summarize conclusions for how the field may attempt to stem future harms caused by CV technologies. We develop narrative case studies to provoke dialogue and deeply explore possible risk scenarios we found to be most probable and severe. We arrive at the position that there are serious potentials for CV to cause discriminatory harm and exacerbate cybersecurity issues.
When creating a new labeled dataset, human analysts or data reductionists must review and annotate large numbers of images. This process is time consuming and a barrier to the deployment of new computervision solutio...
详细信息
ISBN:
(纸本)9781665448994
When creating a new labeled dataset, human analysts or data reductionists must review and annotate large numbers of images. This process is time consuming and a barrier to the deployment of new computervision solutions, particularly for rarely occurring objects. To reduce the number of images requiring human attention, we evaluate the utility of images created from 3D models refined with a generative adversarial network to select confidence thresholds that significantly reduce false alarms rates. The resulting approach has been demonstrated to cut the number of images needing to be reviewed by 50% while preserving a 95% recall rate, with only 6 labeled examples of the target.
This paper presents a new visual motion cue, we call the Visual Threat Cue (VTC) that provides some measure for a relative change in range as well as clearance between a 3D surface and a fixating observer in motion. T...
详细信息
ISBN:
(纸本)0818672587
This paper presents a new visual motion cue, we call the Visual Threat Cue (VTC) that provides some measure for a relative change in range as well as clearance between a 3D surface and a fixating observer in motion. The VTC corresponds to visual fields surrounding a moving observer. The fields are time-based imaginary 3-D surfaces that move with the observer. They are analogous to equi-potential fields of an electric dipole. A practical method to extract the VTC is presented. The approach is independent of the 3D surface texture and needs no optical flow information, 3D reconstruction, segmentation, feature tracking or pre-processing. This algorithm to extract the VTC was applied to several indoor as well as outdoor real images of textures, where we observed a similar behavior for most of the textures employed.
We offer a novel strategy to adapt the perceptual organization process to an object and its context in a scene. Given a set of training images of an object in context, a learning process decides on the relative import...
详细信息
ISBN:
(纸本)0818684976
We offer a novel strategy to adapt the perceptual organization process to an object and its context in a scene. Given a set of training images of an object in context, a learning process decides on the relative importance of the basic Gestalt relationships such as proximity, parallelness, similarity. symmetry, closure, and common region towards segregating the object from the background. This learning is accomplished using a team of stochastic automata in a N-player cooperative game framework. The grouping process which is based on graph partitioning is able to form large groups from relationships defined over a small set of primitives and is fast. We demonstrate the robust performance of the grouping system on a variety of real images. Among the interesting conclusions is the significant role of photometric attributes in grouping and the ability to perform figure-ground segmentation from a set of local relations, each defined over a small number of primitives.
This paper describes an in-depth investigation and implementation of interleaved memory for pixel lookup operations in computervision. Pixel lookup, mapping between coordinates and pixels, is a common operation in co...
详细信息
ISBN:
(纸本)9781424423392
This paper describes an in-depth investigation and implementation of interleaved memory for pixel lookup operations in computervision. Pixel lookup, mapping between coordinates and pixels, is a common operation in computervision, but is also a potential bottleneck due to formidable bandwidth requirements for real-time operation. We focus on the acceleration of pixel lookup operations through parallelizing memory banks by interleaving. The key to applying interleaving for pixel lookup is 2D block data partitioning and support for unaligned access. With this optimization of interleaving, pixel lookup operations can output a block of pixels at once without major overhead for unaligned access. An example implementation of our optimized interleaved memory for affine motion tracking shows that the pixel lookup operations can achieve 12.8 Gbps for random lookup of a 4x4 size block, of 8-bit pixels under 100 MHz operation. Interleaving can be a cost-effective solution for fast pixel lookup in embedded computervision.
We describe a novel algorithm for gait analysis. A person walking frontoparallel to the image plane generates a characteristic 'braided' pattern in a spatiotemporal (XYT) volume. Our algorithm detects this pat...
详细信息
ISBN:
(纸本)0818658274
We describe a novel algorithm for gait analysis. A person walking frontoparallel to the image plane generates a characteristic 'braided' pattern in a spatiotemporal (XYT) volume. Our algorithm detects this pattern, and fits it with a set of spatiotemporal snakes. The snakes can be used to find the bounding contours of the walker. The contours vary over time in a manner characteristic of each walker. Individual gaits can be recognized by applying standard patternrecognition techniques to the contour signals.
In this paper we present a flash game that aims at generating easily ground truth for testing object detection algorithms. Flash the Fish is an online game where the user is shown videos from underwater environments a...
详细信息
ISBN:
(纸本)9780769549903
In this paper we present a flash game that aims at generating easily ground truth for testing object detection algorithms. Flash the Fish is an online game where the user is shown videos from underwater environments and has to take photos of fish by clicking on them. The initial ground truth is provided by object detection algorithms and, subsequent, cluster analysis and user evaluation techniques, allow for the generation of ground truth based on the weighted combination of these "photos". Evaluation of the platform and comparison of the obtained results against a hand drawn ground truth confirmed that reliable ground truth generation is not necessarily a cumbersome task both in terms of effort and time needed.
This paper presents a robust and accurate vision-based augmented reality system for surgical navigation. The key point of our system is a robust and real-time monocular vision algorithm to estimate the 3D pose of surg...
详细信息
This paper presents a robust and accurate vision-based augmented reality system for surgical navigation. The key point of our system is a robust and real-time monocular vision algorithm to estimate the 3D pose of surgical tools, utilizing specially designed code markers and Kalman filter-based position updating. The vision system is not impaired by occlusion and rapid change of illumination. The augmented reality system superimposes the 3D object wireframe onto the live viewing image taken from the surgical microscope as well as displaying other useful navigation information, while allowing the surgeons to freely change its room and focus for viewing. The experimental results verified the robustness and usefulness of the system, and acquired the image registration error less than 2 mm.
暂无评论