Results 1  10
of
13
Contentbased image retrieval at the end of the early years
 IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
, 2000
"... The paper presents a review of 200 references in contentbased image retrieval. The paper starts with discussing the working conditions of contentbased retrieval: patterns of use, types of pictures, the role of semantics, and the sensory gap. Subsequent sections discuss computational steps for imag ..."
Abstract

Cited by 1135 (19 self)
 Add to MetaCart
The paper presents a review of 200 references in contentbased image retrieval. The paper starts with discussing the working conditions of contentbased retrieval: patterns of use, types of pictures, the role of semantics, and the sensory gap. Subsequent sections discuss computational steps for image retrieval systems. Step one of the review is image processing for retrieval sorted by color, texture, and local geometry. Features for retrieval are discussed next, sorted by: accumulative and global features, salient points, object and shape features, signs, and structural combinations thereof. Similarity of pictures and objects in pictures is reviewed for each of the feature types, in close connection to the types and means of feedback the user of the systems is capable of giving by interaction. We briefly discuss aspects of system engineering: databases, system architecture, and evaluation. In the concluding section, we present our view on: the driving force of the field, the heritage from computer vision, the influence on computer vision, the role of similarity and of interaction, the need for databases, the problem of evaluation, and the role of the semantic gap.
A Spatiotemporal Filtering Approach to Motion Segmentation
 Lecture Notes in Computer Science LNCS 2652
, 2003
"... In this paper, a newf requency ased approach to motion segmentation is presented. The proposed technique represents the sequence as a spatiotemporal volume, where a moving o ject corresponds to a threedimensional o ject. In order to detect the "3D volumes" corresponding to significant motions, a ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
In this paper, a newf requency ased approach to motion segmentation is presented. The proposed technique represents the sequence as a spatiotemporal volume, where a moving o ject corresponds to a threedimensional o ject. In order to detect the "3D volumes" corresponding to significant motions, a new scheme ased on a andpass filtering with a setof logGa or spatiotemporal filters is used. It is well known that one of the main pro lems of these approaches is that a filter response varies with the spatial orientation of the underlying signal. To solve this spatial dependency, the proposed model allows to recom ine infb4J185b of motions that has een separated in several filter responses due to its spatial structure. For this purpose, motions are detected as invariance in statistical structure across a rangeof spatiotemporalf requency ands. This technique is illustrated on real and simulated data sets, including sequences with occlusion and transparencies.
A FrequencyDomain Approach for the Extraction of Motion Patterns
 Proceedings of the ICASSP 2003, In press
, 2003
"... In this paper, a new frequencydomain approach to represent motions is proposed. The new scheme is based on a bandpass filtering with a set of logGabor spatiotemporal filters. It is well known that one of the main problems of these approaches is that a filter response varies with the spatial orien ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
In this paper, a new frequencydomain approach to represent motions is proposed. The new scheme is based on a bandpass filtering with a set of logGabor spatiotemporal filters. It is well known that one of the main problems of these approaches is that a filter response varies with the spatial orientation of the underlying signal. To solve this spatial dependency, the proposed model allows to recombine information of motions that has been separated in several filter responses due to its spatial structure. For this purpose, motion patterns are detected as invariance in statistical structure across a range of spatiotemporal frequency bands. This technique is illustrated on real and simulated data sets, including sequences with occlusion and transparencies.
Optical Flow Estimation Based on the Extraction of Motion Patterns
 in Proc IEEE ICIP
, 2003
"... In this paper, a new methodology for optical flow estimation that is able to represent multiple motions is presented. To separate motions at the same location, a new frequencydomain approach is used. This model, based on a bandpass filtering with a set of logGabor spatiotemporal filters, groups t ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
In this paper, a new methodology for optical flow estimation that is able to represent multiple motions is presented. To separate motions at the same location, a new frequencydomain approach is used. This model, based on a bandpass filtering with a set of logGabor spatiotemporal filters, groups together filter responses with continuity in its motion (each group will define a motion pattern). Given a motion pattern, the gradient constraints is applied to the output of each filter in order to obtain multiple estimates of the velocity at the same location. Then, the velocities at each point of the motion pattern are combined using probabilistic rules. The use of "motion patterns" allows to represent multiple motions, while the combination of estimates from different filters helps to reduce the initial aperture problem. This technique is illustrated on real and simulated data sets, including sequences with occlusion and transparencies.
Image representational model for predicting visual distinctness of objects
"... Here we show that a notion of congruence in statistical structure across 20 frequency bands produces a useful definition of visual patterns for perceiving target distinctness. In order to reach such a conclusion,,firstly, the notion of congruence is used to induce a representational model for 20 im ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
Here we show that a notion of congruence in statistical structure across 20 frequency bands produces a useful definition of visual patterns for perceiving target distinctness. In order to reach such a conclusion,,firstly, the notion of congruence is used to induce a representational model for 20 images. Secondly, the visualpattern based representational model is used to define a visual target distinctness metric that involves applying a simple decision rule over the distances between the visual patterns. Finally, a relation is established between the computational distinctness metric and psychophysical target distinctness estimates. 1.
Defining a target distinctness measure through a singlechannel computational model of vision
 Pattern Recognition Letters
, 2003
"... ..."
Best Achievable Compression Ratio for Lossy Image Coding
, 1999
"... High compression ratios are possible at the cost of imperfect source representation, and therefore the natural question is: how much delity in the representation are willing to give up in order to reduce the storage or the number of bits required to transmit the data? The tradeo between image delit ..."
Abstract
 Add to MetaCart
High compression ratios are possible at the cost of imperfect source representation, and therefore the natural question is: how much delity in the representation are willing to give up in order to reduce the storage or the number of bits required to transmit the data? The tradeo between image delity and coding rate is reached with several techniques, but all of them require an ability to measure distortion. The problem is that nding a general enough measure of perceptual quality has proven to be an elusive goal. Here we show how to circumvent the lack of knowledge of what distortion measures are more suitable for images, based on the relationship between information theory and statistics. Given a set of discrete probability distributions, fP q(1) ; P q(2) ; ; P q(K) g, characterizing statistically the reconstructions of the original image at various compression ratios q(1); q(2); ; q(K), we have that each level of compression q(i) divides the set of distributions in tw...
Comparison Of Lossy Coders For Still Images.
"... A criterion for selecting a lossy coder for still images is proposed. The "optimum" coder using this criterion is selected to maximize the minimum Chernoff information between pairs of distributions which are the most difficult to distinguish from achievable and nonachievable regions using this cod ..."
Abstract
 Add to MetaCart
A criterion for selecting a lossy coder for still images is proposed. The "optimum" coder using this criterion is selected to maximize the minimum Chernoff information between pairs of distributions which are the most difficult to distinguish from achievable and nonachievable regions using this coder. The resulting coder has the property that the average Bayesian error probability, for the problems of hypothesis testing between distributions from achievable and nonachievable regions, is less than the average Bayesian error probability for any other choice of coder. The coder selection procedure may be applied without a knowledge of what distortion measure is more suitable for images and does not require an explicit evaluation of error probabilities. Several examples of coder selection are included to illustrate the procedure.
Coder Selection For Lossy Compression Of Biomedical Images.
"... A criterion for selecting lossy coders deemed good enough to submit to clinical evaluation is proposed. This approach does not require an ability to measure distortion or a knowledge of the properties of the human visual perception. The coder selection procedure consists of three steps. For each los ..."
Abstract
 Add to MetaCart
A criterion for selecting lossy coders deemed good enough to submit to clinical evaluation is proposed. This approach does not require an ability to measure distortion or a knowledge of the properties of the human visual perception. The coder selection procedure consists of three steps. For each lossy coder in the class being compared, we first find those levels of compression which are to be the only natural solutions for the tradeoff between source fidelity and coding rate. The second step of the coder selection procedure consists of finding the "optimum" achievable region for each coder, where the optimum achievable region has the property that the average Bayesian error probability for the problems of hypothesis testing between probability distributions associated with decoded images at compression ratios from achievable and nonachievable regions, is less than the average Bayesian error probability for any other choice of achievable region. In the third step of the procedure, a lo...
Axiomatic Characterization Of Relative Information For Predicting Visual Target Distinctness
, 1999
"... Target saliency for humans performing visual search and detection tasks can be estimated by means of the difference between the image from the targetandbackground scene and the image from the same background with no target. It often happens that the structure of these scenes cannot be determined e ..."
Abstract
 Add to MetaCart
Target saliency for humans performing visual search and detection tasks can be estimated by means of the difference between the image from the targetandbackground scene and the image from the same background with no target. It often happens that the structure of these scenes cannot be determined exactly, and therefore the structure of the target image and the image with no target may be characterized statistically by discrete probability distributions. Here we ask the question: What is the amount of relative information between their respective distributions? Due to the availability of a large number of measures, we have to know what postulates and properties should be satisfied by an information theoretic measure in our problem. Thus this paper presents an axiomatic characterization of relative information for the prediction of visual target distinctness from digital imagery. The resultant measures are then applied to quantify the visual distinctness of targets in complex natural back...