Dimensional Histograms Research Articles

General Framework to Reversible Data Hiding for JPEG Images With Multiple Two-Dimensional Histograms

In this paper, a general reversible data hiding (RDH) framework for joint photographic experts group (JPEG) images with multiple two dimensional histograms (2DHs) is proposed. Regardless of whether zero alternating current (AC) coefficients are included to join data embedding or only non-zero AC coefficients are applied, the performance in terms of visual quality and file size increment is improved by using the proposed framework. This framework is mainly composed of the following three parts: histogram generation, adaptive 2DH mapping selection, and improved discrete particle swarm optimization (IDPSO). Unlike existing 2DH-based JPEG RDH methods, in which a uniform threshold is utilized to construct multiple histograms, in histogram generation, thresholds for different histograms are adaptively assigned according to the local properties of histogram coefficients. As a result, as many coefficients in complex regions as possible are excluded from the construction of each histogram. We subtly design multiple 2DH mappings, and adaptively select 2DH mappings for different 2DHs based on their distribution characteristics. Through slight adjustments, each 2DH mapping can be employed in cases where either zero AC coefficients or only non-zero AC coefficients are used for data embedding. Adaptive threshold and 2DH mapping selection provide a better image quality at a given embedding capacity but inevitably cause considerable complexity cost. To significantly reduce the computational cost, we propose IDPSO by combining differential evolution. IDPSO has the advantages of rapid convergence speed as well as satisfactory qualities of the best solutions. With the help of differential evolution, IDPSO expands the diversity of particles and efficiently avoids local optimal trapping problems. The experimental results also demonstrate the effectiveness of the proposed method in terms of visual quality, file size increment and complexity cost.

Optimization of Cr Seed Layer Effect for Surface Roughness of As-Deposited Silver Film using Electron Beam Deposition Method

The surface roughness is an important parameter in determining the physical properties and quality of thin films deposited by physical vapor deposition (PVD) method. The presence of an intermediate layer between metallic nanoparticles and substrate significantly promotes the adhesion and reduces the surface roughness. In this article, we have investigated the effect of Chromium (Cr) seed layer to optimize the surface roughness on the growth of as-deposited silver (Ag) film using borosilicate glass and silicon wafer substrates. For this purpose, Ag thin films were deposited with a Cr seed layer of different thickness on borosilicate glass and silicon wafer substrates using an electron beam (E-Beam) deposition method. The Cr thin film of different thickness ranging from 1 nm to 6 nm was thermally evaporated and pure Ag with the same thickness was evaporated at the same rate on previously coated substrates. The deposition of the nanostructured thin film was confirmed by UV-Vis and XRD characterizations. The difference in transmittance for uncoated and coated substrates ensured the deposition. The presence of pure Ag crystalline phase was confirmed by XRD pattern. Surface roughness was measured using Atomic Force Microscopy (AFM) and the conductance was measured using 4-probe conductivity method. The density of nanoparticles and smoothness were visualized from two dimensional (2D) and three dimensional (3D) surface height histograms of representative AFM images. The quantitative roughness was measured in terms of root mean square (RMS) roughness and mean roughness. The high dense and smoother thin films were found for ~2-4 nm Cr layer thickness in case of the glass substrate. The slight increase in roughness was observed for ~1-6 nm Cr layer thickness in case of the silicon substrate. The dependence of the conductivity of thin films on surface roughness is investigated to verify the effect of surface roughness on different applications of Ag thin film. The conductance results have been analyzed as; for a glass substrate, conductivity was maximum for thin films containing ~2 nm Cr seed layer thickness, while for silicon substrate the maximum conductivity was found for the thin film containing ~1 nm Cr seed layer.

Optimization of Cr Seed Layer Effect for Surface Roughness of As-Deposited Silver Film using Electron Beam Deposition Method

The surface roughness is an important parameter in determining the physical properties and quality of thin films deposited by physical vapor deposition (PVD) method. The presence of an intermediate layer between metallic nanoparticles and substrate significantly promotes the adhesion and reduces the surface roughness. In this article, we have investigated the effect of Chromium (Cr) seed layer to optimize the surface roughness on the growth of as-deposited silver (Ag) film using borosilicate glass and silicon wafer substrates. For this purpose, Ag thin films were deposited with a Cr seed layer of different thickness on borosilicate glass and silicon wafer substrates using an electron beam (E-Beam) deposition method. The Cr thin film of different thickness ranging from 1 nm to 6 nm was thermally evaporated and pure Ag with the same thickness was evaporated at the same rate on previously coated substrates. The deposition of the nanostructured thin film was confirmed by UV-Vis and XRD characterizations. The difference in transmittance for uncoated and coated substrates ensured the deposition. The presence of pure Ag crystalline phase was confirmed by XRD pattern. Surface roughness was measured using Atomic Force Microscopy (AFM) and the conductance was measured using 4-probe conductivity method. The density of nanoparticles and smoothness were visualized from two dimensional (2D) and three dimensional (3D) surface height histograms of representative AFM images. The quantitative roughness was measured in terms of root mean square (RMS) roughness and mean roughness. The high dense and smoother thin films were found for ~2-4 nm Cr layer thickness in case of the glass substrate. The slight increase in roughness was observed for ~1-6 nm Cr layer thickness in case of the silicon substrate. The dependence of the conductivity of thin films on surface roughness is investigated to verify the effect of surface roughness on different applications of Ag thin film. The conductance results have been analyzed as; for a glass substrate, conductivity was maximum for thin films containing ~2 nm Cr seed layer thickness, while for silicon substrate the maximum conductivity was found for the thin film containing ~1 nm Cr seed layer.

Violent scene detection algorithm based on kernel extreme learning machine and three-dimensional histograms of gradient orientation

Most existing feature descriptors for video have limited representation ability. In order to improve the recognition accuracy of method for detecting the videos that include violent scenes and take advantage of the logical structure of video sequences, a novel feature constructing approach based on three dimensional histograms of gradient orientation (HOG3D), the Bag of Visual Words (BoVW) model, and feature pooling technology is proposed. This approach, combined with kernel extreme learning machine (KELM), can be used to detect violent scene. First, the HOG3D feature is extracted on the block level for video, and then the K-Means clustering algorithm is implemented to generate visual words. Then, the bag of visual words framework is used for the quantization of feature. And the feature pooling technology is operated to generate a feature vector for an entire video segment, and feature vectors of training data and testing data were used separately to train the model and evaluate the performance of the proposed approach. The experimental results showed that the proposed feature descriptor had good representation and generalization abilities. The proposed approach is efficient for violent scene detection, and the accuracy matches the best result on Hockey dataset, and it outperforms state-of-the-art on Movies.

Deformation-Electromagnetic Relations in Lithospheric Activity Manifestations

It has been shown that dipole radiation of accelerated charges, described by Larmor relation, is the basis of the known mechanic-electromagnetic processes of rock deformation. Comparison of crust deformation acceleration with natural electromagnetic field parameters of ELF-VLF range showed good relation. It manifests in the maxima of occurrence frequency density of synchronous deformation-electromagnetic events on two dimensional histograms. The data of a laser strain-meter and a recorder of natural electromagnetic radiation of ELF-VLF range, recorded in a zone of increased seismic activity (Kamchatka, Karymshina site), were used. The authors made an assumption on the existence of stationary regions of deformation process and mechanic-electromagnetic transformations corresponding to regions with different mechanic properties and rock petrographic composition.

Development of high resolution phoswich depth-of-interaction block detectors utilizing Mg co-doped new scintillators

To correct for parallax error in positron emission tomography (PET), phoswich depth-of-interaction (DOI) detector using multiple scintillators with different decay times is a practical approach. However not many scintillator combinations suitable for phoswich DOI detector have been reported. Ce doped Gd3Ga3Al2O12 (GFAG) is a newly developed promising scintillator for PET detector, which has high density, high light output, appropriate light emission wavelength for silicon-photomultiplier (Si-PM) and faster decay time than that of Ce doped Gd3Al2Ga3O12 (GAGG). In this study, we developed a Si-PM based phoswich DOI block detector of GFAG with GAGG crystal arrays and evaluated its performance. We assembled a GFAG block and a GAGG block and they were optically coupled in depth direction to form a phoswich detector block. The phoswich block was optically coupled to a Si-PM array with a 1 mm thick light guide. The sizes of the GFAG and GAGG pixels were 0.9 mm x 0.9 mm x 7.5 mm and they were arranged into 24 x 24 matrix with 0.1 mm thick BaSO4 as reflector. We conducted the performance evaluation for two types of configurations; GFAG block arranged in upper layer (GFAG/GAGG) and GAGG arranged in upper layer (GAGG/GFAG). The measured two dimensional position histograms of these block detectors showed good separation and pulse shape spectra produced two distinct peaks for both configurations although some difference in energy spectra were observed. These results indicate phoswich block detectors composed of GFAG and GAGG are promising for high resolution DOI PET systems.

Light wavelength influence on surface plasmon resonance in citrate–gold nanosystems

Citrate–gold particles were yielded according to classical method of auric salt reduction in two different synthesis media aiming to use them further applications in biomedical and environmental domains. The analysis of citrate–gold interaction was done through UV–vis and IR spectroscopy as well as by Transmission Electron Microscopy (TEM) and Dark Field (DF) Microscopy. Average particle size was higher for citrate–gold NPs synthesized with NaOH (32.5 nm) than for NPs synthesized with NaCl (15 nm). Dimensional histograms of one year aged colloidal suspensions presented mean size of 29 nm and respectively 18 nm. The influence of 90 min light exposure, analyzed by UV–vis, evidenced that for both NaOH synthesis protocol and NaCl protocol, plasmon band maxima at 528 nm and respectively 538 nm didn't changed, neither for white nor for green light. For one year aged samples this band shifted to 540 nm for green light irradiation in the case of citrate–gold NPs synthesized with NaOH. Also, for these NPs, both green and white light exposures resulted in plasmon band intensity changes for native as well as for aged samples. FTIR investigation showed also different changes at the level of the intensity of main vibration bands of citrate–gold after exposure to light, suggesting stronger adsorption of citrate in the case of NaCl addition in the initial reaction medium than in the case of NaOH. Finally, the utilization of NaCl in the synthesis protocol seems to favor the synthesis of more stable and lower toxicity colloidal suspensions, both during time and under the light irradiation.

Precursor configurations and post-rupture evolution of Ag-CO-Ag single-molecule junctions.

Experimental correlation analysis and first-principles theory are used to probe the structure and evolution of Ag-CO-Ag single-molecule junctions both before the formation and after the rupture of the junctions. Two dimensional correlation histograms and conditional histograms demonstrate that prior to the single-molecule bridge configuration the CO molecule is already bound parallel to the Ag single-atom contact. This molecular precursor configuration is accompanied by the opening of additional conductance channels compared to the single-channel transport in pure Ag monoatomic junctions. To investigate the post-rupture evolution of the junction we introduce a cross-correlation analysis between the opening and the subsequent closing conductance traces. This analysis implies that the molecule is bound rigidly to the apex of one electrode, and so the same single-molecule configuration is re-established as the junction is closed. The experimental results are confirmed by ab initio simulations of the evolution of contact geometries, transmission eigenvalues and scattering wavefunctions.

DPSynthesizer: Differentially Private Data Synthesizer for Privacy Preserving Data Sharing.

Differential privacy has recently emerged in private statistical data release as one of the strongest privacy guarantees. Releasing synthetic data that mimic original data with Differential privacy provides a promising way for privacy preserving data sharing and analytics while providing a rigorous privacy guarantee. However, to this date there is no open-source tools that allow users to generate differentially private synthetic data, in particular, for high dimensional and large domain data. Most of the existing techniques that generate differentially private histograms or synthetic data only work well for single dimensional or low-dimensional histograms. They become problematic for high dimensional and large domain data due to increased perturbation error and computation complexity. We propose DPSynthesizer, a toolkit for differentially private data synthesization. The core of DPSynthesizer is DPCopula designed for high-dimensional and large-domain data. DPCopula computes a differentially private copula function from which synthetic data can be sampled. Copula functions are used to describe the dependence between multivariate random vectors and allow us to build the multivariate joint distribution using one-dimensional marginal distributions. DPSynthesizer also implements a set of state-of-the-art methods for building differentially private histograms, suitable for low-dimensional data, from which synthetic data can be generated. We will demonstrate the system using DPCopula as well as other methods with various data sets and show the feasibility, utility, and efficiency of various methods.

Timing, polarimetry and physics of the bright, nearby millisecond pulsar PSR J0437−4715 – a single-pulse perspective

Single pulses from radio pulsars contain a wealth of information about emission and propagation in the magnetosphere and insight into their timing properties. It was recently demonstrated that single-pulse emission is responsible for limiting the timing stability of the brightest of millisecond pulsars. We report on an analysis of more than a million single-pulses from PSR J0437-4715 and present various statistical properties such as the signal-to-noise ratio (S/N) distribution, timing and polarimetry of average profiles integrated from subpulses with chosen S/N cut-offs, modulation properties of the emission, phase-resolved statistics of the S/N, and two dimensional spherical histograms of the polarization vector orientation. The last of these indicates the presence of orthogonally polarised modes (OPMs). Combined with the dependence of the polarisation fraction on the S/N and polarimetry of the brightest pulses, the existence of OPMs constrains pulsar emission mechanisms and models for the plasma physics in the magnetosphere.

Differentially Private Synthesization of Multi-Dimensional Data using Copula Functions.

Differential privacy has recently emerged in private statistical data release as one of the strongest privacy guarantees. Most of the existing techniques that generate differentially private histograms or synthetic data only work well for single dimensional or low-dimensional histograms. They become problematic for high dimensional and large domain data due to increased perturbation error and computation complexity. In this paper, we propose DPCopula, a differentially private data synthesization technique using Copula functions for multi-dimensional data. The core of our method is to compute a differentially private copula function from which we can sample synthetic data. Copula functions are used to describe the dependence between multivariate random vectors and allow us to build the multivariate joint distribution using one-dimensional marginal distributions. We present two methods for estimating the parameters of the copula functions with differential privacy: maximum likelihood estimation and Kendall's τ estimation. We present formal proofs for the privacy guarantee as well as the convergence property of our methods. Extensive experiments using both real datasets and synthetic datasets demonstrate that DPCopula generates highly accurate synthetic multi-dimensional data with significantly better utility than state-of-the-art techniques.

SLUG—STOCHASTICALLY LIGHTING UP GALAXIES. I. METHODS AND VALIDATING TESTS

The effects of stochasticity on the luminosities of stellar populations are an often neglected but crucial element for understanding populations in the low mass or low star formation rate regime. To address this issue, we present SLUG, a new code to "Stochastically Light Up Galaxies". SLUG synthesizes stellar populations using a Monte Carlo technique that treats stochastic sampling properly including the effects of clustering, the stellar initial mass function, star formation history, stellar evolution, and cluster disruption. This code produces many useful outputs, including i) catalogs of star clusters and their properties, such as their stellar initial mass distributions and their photometric properties in a variety of filters, ii) two dimensional histograms of color-magnitude diagrams of every star in the simulation, iii) and the photometric properties of field stars and the integrated photometry of the entire simulated galaxy. After presenting the SLUG algorithm in detail, we validate the code through comparisons with starburst99 in the well-sampled regime, and with observed photometry of Milky Way clusters. Finally, we demonstrate the SLUG's capabilities by presenting outputs in the stochastic regime. SLUG is publicly distributed through the website http://sites.google.com/site/runslug/.

Multiparameter Fluorescence Image Spectroscopy

In multiparameter fluorescence image spectroscopy (MFIS) simultaneously all information observable in a confocal fluorescence experiment is recorded photon by photon. Subsequent analysis of the primary fluorescence data (polarization, color, arrival time and spatial origin of the photons) enables e.g. calculation of fluorescence correlation / cross correlation spectra (FCS / FCCS), Förster resonance energy transfer (FRET) efficiency, fluorescence lifetime and anisotropy. Display of the results in more dimensional histograms to combine the fluorescence parameters of interest leads to species selective images and thus maximizes the selectivity of fluorescence microscopic methods.

MCNP output data analysis with ROOT (MODAR)

MCNP Output Data Analysis with ROOT (MODAR) is a tool based on CERN's ROOT software. MODAR has been designed to handle time-energy data issued by MCNP simulations of neutron inspection devices using the associated particle technique. MODAR exploits ROOT's Graphical User Interface and functionalities to visualize and process MCNP simulation results in a fast and user-friendly way. MODAR allows to take into account the detection system time resolution (which is not possible with MCNP) as well as detectors energy response function and counting statistics in a straightforward way. New version program summary Program title: MODAR Catalogue identifier: AEGA_v1_1 Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEGA_v1_1.html Program obtainable from: CPC Program Library, Queen's University, Belfast, N. Ireland Licensing provisions: Standard CPC licence, http://cpc.cs.qub.ac.uk/licence/licence.html No. of lines in distributed program, including test data, etc.: 150 927 No. of bytes in distributed program, including test data, etc.: 4 981 633 Distribution format: tar.gz Programming language: C++ Computer: Most Unix workstations and PCs Operating system: Most Unix systems, Linux and windows, provided the ROOT package has been installed. Examples where tested under Suse Linux and Windows XP. RAM: Depends on the size of the MCNP output file. The example presented in the article, which involves three two dimensional 139 × 740 bins histograms, allocates about 60 MB. These data are running under ROOT and include consumption by ROOT itself. Classification: 17.6 Catalogue identifier of previous version: AEGA_v1_0 Journal reference of previous version: Comput. Phys. Comm. 181 (2010) 1161 External routines: ROOT version 5.24.00 ( http://root.cern.ch/drupal/) Does the new version supersede the previous version?: Yes Nature of problem: The output of a MCNP simulation is an ascii file. The data processing is usually performed by copying and pasting the relevant parts of the ascii file into Microsoft Excel. Such an approach is satisfactory when the quantity of data is small but is not efficient when the size of the simulated data is large, for example when time-energy correlations are studied in detail such as in problems involving the associated particle technique. In addition, since the finite time resolution of the simulated detector cannot be modeled with MCNP, systems in which time-energy correlation is crucial cannot be described in a satisfactory way. Finally, realistic particle energy deposit in detectors is calculated with MCNP in a two step process involving type-5 then type-8 tallies. In the first step, the photon flux energy spectrum associated to a time region is selected and serves as a source energy distribution for the second step. Thus, several files must be manipulated before getting the result, which can be time consuming if one needs to study several time regions or different detectors performances. In the same way, modeling counting statistics obtained in a limited acquisition time requires several steps and can also be time consuming. Solution method: In order to overcome the previous limitations, the MODAR C++ code has been written to make use of CERN's ROOT data analysis software. MCNP output data are read from the MCNP output file with dedicated routines. Two dimensional histograms are filled and can be handled efficiently within the ROOT framework. To keep a user friendly analysis tool, all processing and data display can be done by means of ROOT Graphical User Interface. Specific routines have been written to include detectors finite time resolution and energy response function as well as counting statistics in a straightforward way. Reasons for new version: For applications involving the Associate Particle Technique, a large number of gamma rays are produced by the fast neutrons interactions. To study the energy spectra, it is useful to identify the gamma-ray energy peaks in a straightforward way. Therefore, the possibility to show gamma rays corresponding to specific reactions has been added in MODAR. Summary of revisions: It is possible to use a gamma ray database to better identify in the energy spectra gamma ray peaks with their first and second escapes. Histograms can be scaled by the number of source particle to evaluate the number of counts that is expected without statistical uncertainties. Additional comments: The possibility of adding tallies has also been incorporated in MODAR in order to describe systems in which the signal from several detectors can be summed. Moreover, MODAR can be adapted to handle other problems involving two dimensional data. Running time: The CPU time needed to smear a two dimensional histogram depends on the size of the histogram. In the presented example, the time-energy smearing of one of the 139 × 740 two dimensional histograms takes 3 minutes with a DELL computer equipped with INTEL Core 2.

A coarse-to-fine framework to efficiently thwart plagiarism

This paper presents a systematic framework using multilevel matching approach for plagiarism detection (PD). A multilevel structure, i.e. document–paragraph–sentence, is used to represent each document. In document and paragraph level, we use traditional dimensionality reduction technique to project high dimensional histograms into latent semantic space. The Earth Mover’s Distance (EMD), instead of exhaustive matching, is employed to retrieve relevant documents, which enables us to markedly shrink the searching domain. Two PD algorithms are designed and implemented to efficiently flag the suspected plagiarized document sources. We conduct extensive experimental verifications including document retrieval, PD, the study of the effects of parameters, and the empirical study of the system response. The results corroborate that the proposed approach is accurate and computationally efficient for performing PD.

Choice of a pertinent color space for color texture characterization using parametric spectral analysis

This article presents a comparison of different color spaces including RGB, IHLS and L⁎a*b* for color texture characterization. This comparison is based on the fusion of the independent spatial structure and color feature cues. In IHLS and L*a*b*, two channel complex color images are created from the luminance and the chrominance values. For such images, two dimensional complex multichannel linear prediction models are used to perform parametric power spectrum estimation and the structure feature cues are computed from this estimated power spectrum. Quantitative comparison of auto spectra of luminance and combined chrominance channels for different color spaces is done. This comparison is based on the degree of decorrelation between luminance and chrominance information provided by different color space transformations. Three dimensional histograms are used as color feature cues. Then, to classify color textures, Kullback–Leibler divergence based symmetric distance measures are calculated for pure color, luminance structure and chrominance structure feature cues. Individual as well as combined effect of information from all feature cues on classification results is then compared for different color spaces and different color texture data sets. The proposed color texture classification method performs better than the state of the art methods in certain cases. The L*a*b* color space gives us a better characterization of the chrominance spatial structure as well as the overall spatial structure for all of the chosen data sets. Experimental results on pixel classification of color textures are also presented and discussed.

Pitch-frequency histogram-based music information retrieval for Turkish music

This study reviews the use of pitch histograms in music information retrieval studies for western and non-western music. The problems in applying the pitch-class histogram-based methods developed for western music to non-western music and specifically to Turkish music are discussed in detail. The main problems are the assumptions used to reduce the dimension of the pitch histogram space, such as, mapping to a low and fixed dimensional pitch-class space, the hard-coded use of western music theory, the use of the standard diapason (A4=440 Hz), analysis based on tonality and tempered tuning. We argue that it is more appropriate to use higher dimensional pitch-frequency histograms without such assumptions for Turkish music. We show in two applications, automatic tonic detection and makam recognition, that high dimensional pitch-frequency histogram representations can be successfully used in Music Information Retrieval (MIR) applications without such pre-assumptions, using the data-driven models.

Count Method of Leaves on the Surface of a River Using Image Processing

This paper proposes a new counting method for floating leaves on the surface of a river to be used in ecosystem monitoring. Since conventional counting methods (such as the litter trap method) require considerable manual labor for precise monitoring of material flow in ecosystem, an efficient counting method was needed. Our method automatically counts the number of floating leaves in recorded video images. Floating leaves are detected using color and motion features. The color feature is represented by 3 dimensional histograms of the RGB color space. For the motion feature, speed and acceleration of the targets are used. The counting method proposed in this paper has been applied to a five hours video which recorded 20,000 leaves, and high recall and precision rates of 96% and 94%, respectively, have been achieved. This paper also proposes a user interface for training the counting mechanism based on the Interactive Machine Learning model. The user can easily produce a huge number of sample data to train the detection mechanism by using the user interface in the same way as coloring a picture.

Poster - Thurs Eve-01: Comparison of clinical IMRT plan quality and delivery accuracy: Few large segments vs many small segments.

Commercial radiation treatment planning systems for intensity modulation use optimization algorithms that can vary multi-leaf collimator (MLC) segment sizes, segment number and the minimum number of monitor units (MU) per segment. These parameters are varied according to the treatment site, size, location, and proximity to the organs at risk. This study compares the utility of optimization using (Case A) few large segments and a higher minimum MU per segment to that of (case B) using many smaller segments with a lower minimum MU per segment. For Case A, the patient benefits from a reduced treatment time associated with fewer MUs and fewer MLC movements and an increased accuracy in dose delivery. Also, shorter treatment times may lead to fewer patient movement uncertainties. The accumulated MLC leakage dose is reduced, the patient specific quality assurance (QA) is more manageable and small field modeling inaccuracies are reduced. Pinnacle-3 (v8) plans are generated with direct machine parameter optimization (DMPO) for both scenarios. Three dimensional dose distributions and dose volume histograms are used to compare plan quality. We compare plans using few large MLC segments with those using many small MLC segments for some clinical cases. Improved plan quality is demonstrated using fewer MLC segments. Dose QAs are performed and compared for each scenario using MapCheck and film. When comparing dose delivery accuracy between different MU per segment settings, a decrease in delivery errors with minimum MU size is observed. In conclusion, few large MLC segments with larger area should be used when possible.

Generalizing the Lucas–Kanade algorithm for histogram-based tracking

We present a novel histogram-based tracking algorithm, which is a generalization of the template matching Lucas–Kanade algorithm (and in particular of the inverse compositional version which is more efficient). The algorithm does not make use of any spatial kernel. Instead, the dependency of the histogram on the warping parameters is introduced via a feature kernel. This fact helps us to overcome several limitations of kernel-based methods. The target is represented by a collection of patch-based histograms, thus retaining spatial information. A robust statistics scheme assigns weights to the different patches, rendering the algorithm robust to partial occlusions and appearance changes. We present the algorithm for 1-D histograms (e.g. gray-scale), however extending the algorithm to handle higher dimensional histograms (e.g. color) is straightforward. Our method applies to any warping transformation that forms a group, and to any smooth feature. It has the same asymptotic complexity as the original inverse compositional template matching algorithm. We present experimental results which demonstrate the robustness of our algorithm, using only gray-scale histograms.