Search

We prove new explicit asymptotic formulae between (geometric) eigenvalues in max-algebra and classical distinguished eigenvalues of nonnegative matrices, which are useful tools for transferring results between both settings. We establish new inequalities for both types of eigenvalues of Hadamard products and Hadamard weighted geometric means of nonnegative matrices. Moreover, a version of the spectral mapping theorem for the distinguished spectrum is pointed out.

In this paper, we determine periodicity of quantum walks defined by mixed paths and mixed cycles. By the spectral mapping theorem of quantum walks, consideration of periodicity is reduced to eigenvalue analysis of $\eta$-Hermitian adjacency matrices. First, we investigate coefficients of the characteristic polynomials of $\eta$-Hermitian adjacency matrices. We show that the characteristic polynomials of mixed trees and their underlying graphs are same. We also define $n+1$ types of mixed cycles and show that every mixed cycle is switching equivalent to one of them. We use these results to discuss periodicity. We show that the mixed paths are periodic for any $\eta$. In addition, we provide a necessary and sufficient condition for a mixed cycle to be periodic and determine their periods., Comment: 22 pages, 12 figures

We present a database of parallel recordings of speech and singing, collected and released by the Human Language Technology (HLT) laboratory at the National University of Singapore (NUS), that is called NUS-HLT Speak-Sing (NHSS) database. We release this database to the public to support research activities, that include, but not limited to comparative studies of acoustic attributes of speech and singing signals, cooperative synthesis of speech and singing voices, and speech-to-singing conversion. This database consists of recordings of sung vocals of English pop songs, the spoken counterpart of lyrics of the songs read by the singers in their natural reading manner, and manually prepared utterance-level and word-level annotations. The audio recordings in the NHSS database correspond to 100 songs sung and spoken by 10 singers, resulting in a total of 7 hours of audio data. There are 5 male and 5 female singers, singing and reading the lyrics of 10 songs each. In this paper, we discuss the design methodology of the database, analyse the similarities and dissimilarities in characteristics of speech and singing voices, and provide some strategies to address relationships between these characteristics for converting one to another. We develop benchmark systems, which can be used as reference for speech-to-singing alignment, spectral mapping, and conversion using the NHSS database., Accepted to Speech Communication

Let $$(T(t))_{t\ge 0}$$ be a strongly continuous semigroup of operators on a Banach space X and A be its generator. We show, in this paper, the following results: $$\sigma _{*}(T(t))\setminus \{0\}=\exp (t\sigma _{*}(A)),$$ for $$t>0$$ , where $$\sigma _{*}$$ runs over a family of some usual spectra.

This paper is devoted to the study of some new definition of pseudo-spectrum for a matrix pencils called generalized determinant pseudo-spectrum. Some properties of the generalized determinant pseudo-spectrum are investigated and some examples are also given. Finally, we show an analogue of the spectral mapping theorem for the generalized determinant pseudo-spectrum in the matrix algebra.

In this paper we study the approximate point spectrum and the surjectivity spectrum of a Toeplitz operators $$T_\phi$$ , defined on Hardy spaces $$H^2(\mathbf{T })$$ of the unit circle $$\mathbf{T }$$ of $${\mathbb {C}}$$ , where the symbol $$\phi \in L^\infty (\mathbf{T })$$ , or $$\phi$$ is a continuous symbol. We also extend some results obtained by Farenick and Lee (Trans Am Math Soc 348(10):4153–4174, 1996). In particular, we study Weyl type theorems and the spectral mapping theorem for the Weyl spectra.

We generalize the classical Riesz–Dunford holomorphic functional calculus with weakly locally nilpotent operators in a Frechet space to the case of the algebra of all formal power series. We obtain a counterpart of the spectral mapping theorem and formulas similar to the Taylor and Riesz–Dunford formulas. We show that the weak local nilpotency of an operator is necessary for constructing the formal functional calculus. Bibliography: 12 titles.

Detailed investigation on texture and mineralogy of beach sediments helps to understand their nature of deposition and potential targets of strategic mineral deposits. The beach sediments from the ...

The deposition of amyloid plaques is considered one of the main microscopic features of Alzheimer's disease (AD). Since plaque formation can precede extensive neurodegeneration and it is the main clinical manifestation of AD, it constitutes a relevant target for new treatment and diagnostic approaches. Micro-Raman spectroscopy, a label-free technique, is an accurate method for amyloid plaque identification and characterization. Here, we present a high spatial resolution micro-Raman hyperspectral study in transgenic APPswePS1ΔE9 mouse brains, showing details of AD tissue biochemical and histological changes without staining. First we used stimulated micro-Raman scattering to identify the lipid-rich halo surrounding the amyloid plaque, and then proceeded with spontaneous (conventional) micro-Raman spectral mapping, which shows a cholesterol and sphingomyelin lipid-rich halo structure around dense-core amyloid plaques. The detailed images of this lipid halo relate morphologically well with dystrophic neurites surrounding plaques. Principal Component Analysis (PCA) of the micro-Raman hyperspectral data indicates the feasibility of the optical biomarkers of AD progression with the potential for discriminating transgenic groups of young adult mice (6-month-old) from older ones (12-month-old). Frequency-specific PCA suggests that plaque-related neurodegeneration is the predominant change captured by Raman spectroscopy, and the main differences are highlighted by vibrational modes associated with cholesterol located majorly in the lipid halo.

Let $$(T(t))_{t\ge 0}$$ be a $$C_0$$ semigroup on a Banach space X with infinitesimal generator A. In this work, we give conditions for which the spectral mapping theorem $$\sigma _{*}(T(t))\backslash \{0\}=\{e^{\lambda s}, \lambda \in \sigma _{*}(A)\}$$ holds, where $$\sigma _*$$ can be equal to the essential, Browder and Kato spectrum. Also, we will be interested in the relations between the spectrum of A and the spectrum of the $$n^{th}$$ derivative $$T(t)^{(n)}$$ of a differentiable $$C_0$$ semigroup $$(T(t))_{t\ge 0}$$ .

This study deals with sound radiation from a thin clamped annular plate. The fluid loading effect is included. Although it is a classical problem, using direct numerical integration to solve such a problem is usually troublesome and time consuming since the coupling integrals have certain singularities. Therefore, this study applies radial polynomials to improve the time efficiency of numerical calculations with no loss of accuracy. As radial polynomials are usually used for circular geometry, spectral mapping has been proposed to perform the necessary conversion from circular to annular geometry. The conversion is performed in the wavenumber domain. Spectral mapping makes it possible to efficiently obtain a number of numerical results for the acoustic power and acoustic field generated.

In this article, we have processed Landsat-8 spectral bands, ground gravity, and magnetic data for delineating the mafic cumulates within the anorthositic-layered complex in Sittampundi, Tamilnadu,...

This chapter first reviews recent pilot studies covering limited areas often with exposed bare soils. Then we focus on the challenges for large-scale application of spectral mapping when the soil and parent material are heterogeneous and surface conditions are unknown. In order to deal with these constraints we discuss i) calibration of spectral models based on large spectral libraries, ii) surface conditions that disturb the soil signal, and iii) time series of images in order to delimit cropland fields and increase the extent of bare soil that can be mapped. Finally, a case study deals with a SOC prediction map derived from the spectra of a Sentinel-2 image and calibrated using the LUCAS spectral library.

In this paper, we propose a novel approach for a voice conversion system that makes effective use of spectral characteristics and excitation information, to optimally morph voice. This work addresses some key issues that are not adequately addressed in reported literature and achieves a more holistic voice conversion system. This is achieved using a strategic combination of line spectral frequencies (LSFs) to minimize the effects of over smoothing, a neural network for performing nonlinear spectral mapping and time-domain pitch synchronous overlap add to account for the interaction of excitation signal with the vocal tract. Within this proposed system, two different methods of pitch modification have been suggested, and the performance of these is compared with existing models of comparable complexity. The proposed methods have an average LSF performance index of 0.4082 and 0.4008, respectively, which is higher than existing similar work reported.

In this paper, we present a novel framework for a voice conversion (VC) system based on a cyclic recurrent neural network (CycleRNN) and a finely tuned WaveNet vocoder. Even though WaveNet is capable of producing natural speech waveforms when fed with natural speech features, it still suffers from speech quality degradation when fed with oversmoothed features, such as spectral parameters estimated from a statistical model. One way to address this problem is to introduce oversmoothed features while developing a WaveNet model. However, in a VC framework, providing oversmoothed spectral features of a target speaker for WaveNet modeling is not straightforward owing to the difference in the time-sequence alignment from that of a source speaker. To overcome this problem, we propose the use of a cyclic spectral conversion network, i.e., CycleRNN, capable of performing a conversion flow, i.e., source-to-target, and a cyclic flow, i.e., to generate self-predicted target spectra. The CycleRNN spectral model is trained using both conversion and weighted cyclic losses. To finely tune WaveNet, a pretrained multispeaker WaveNet model is optimized using the self-predicted features of the corresponding target speaker of a speaker conversion pair. The experimental results demonstrate that 1) the proposed CycleRNN-based spectral model for WaveNet fine-tuning significantly improves the naturalness of the converted speech waveforms, giving an overall mean opinion score of 3.50; and 2) the proposed model yields the highest speaker conversion accuracy, giving an overall speaker similarity score of 78.33%, which is a significant improvement compared with conventional WaveNet fine-tuning using natural target features.

In this paper, we consider the asymptotic behavior of solutions to the linear spatially homogeneous Boltzmann equation for hard potentials without angular cutoff. We obtain an optimal rate of exponential convergence towards equilibrium in a L1-space with a polynomial weight. Our strategy is taking advantage of a spectral gap estimate in the Hilbert space L 2 ( μ − 1 2 ) and a quantitative spectral mapping theorem developed by Gualdani et al. (2017).

Herein we report the detection and differentiation of plasmonic and non-plasmonic nanoparticles simultaneously administered to A549 lung epithelium cells using dark-field microscopy and hyperspectral imaging. Reflectance spectra-based hyperspectral mapping and image analysis allows for the effective quasi-quantitative identification of nanomaterials in cultured human cells. © 2018 The Chemical Society of Japan. The work is performed according to the Russian Government Program of Competitive Growth of Kazan Federal University and funded by Russian presidential grant MD-6655.2018.4 and by the subsidy allocated to Kazan Federal University for the state assignment in the sphere of scientific activities (6.7743.2017/6.7).

The behavior of the domain, the range, the kernel and the multi-valued part of a rational function in a linear relation is analyzed, respectively. We give some basic properties of such linear relations, and we prove that the rational form of the spectral mapping theorem holds in terms of ascent, essential ascent, descent and essential descent.

In this paper, we study properties of the operator equation TT*=T+T* which T.T. West observed in [12]. We first investigate the structure of solutions T 2 B(H) of such equation. Moreover, we prove that if T is a polynomial root of solutions of that operator equation, then the spectral mapping theorem holds for Weyl and essential approximate point spectra of T and f(T) satisfies a-Weyl?s theorem for f?H(?(T)), where H(?(T)) is the space of functions analytic in an open neighborhood of ?(T).

We extend and deepen the theory of functional calculus for semigroup generators, based on the algebra $\mathcal B$ of analytic Besov functions, which we initiated in a previous paper. In particular, we show that our construction of the calculus is optimal in several natural senses. Moreover, we clarify the structure of $\mathcal B$ and identify several important subspaces in practical terms. This leads to new spectral mapping theorems for operator semigroups and to wide generalisations of a number of basic results from semigroup theory., 51 pages. This is a version of the paper to appear in Journal of Functional Analysis

Chemical analysis is commonly used in the field of forensic science where the precise discrimination of primary evidence is of significant importance. Laser-Induced Breakdown Spectroscopy (LIBS) exceeds other spectroscopic methods in terms of the time required for pre- and post-sample preparation, the insensitivity to sample phase state be it solid, liquid, or gas, and the detection of two-dimensional spectral mapping from real time point measurements. In this research, fingerprint samples on various surface materials are considered in the chemical detection and reconstruction of fingerprints using the two-dimensional LIBS technique. Strong and distinct intensities of specific wavelengths represent visible ink, natural secretion of sweat, and contaminants from the environment, all of which can be present in latent fingerprints. The particular aim of the work presented here is to enhance the precision of the two-dimensional recreation of the fingerprints present on metal, plastic, and artificially prepared soil surface using LIBS with principal component analysis. By applying a distinct wavelength discrimination for two overlapping fingerprint samples, separation into two non-identical chemical fingerprints was successfully performed.

In this paper, we consider the quantum walk on $\mathbb{Z}$ with attachment of one-length path periodically. This small modification to $\mathbb{Z}$ provides localization of the quantum walk. The eigenspace causing this localization is generated by finite length round trip paths. We find that the localization is due to the eigenvalues of an underlying random walk. Moreover we find that the transience of the underlying random walk provides a slow down of the pseudo velocity of the induced quantum walk and a different limit distribution from the Konno distribution.

This paper studies the spectrum of a multi-dimensional split-step quantum walk with a defect that cannot be analysed in the previous papers. To this end, we have developed a new technique which allow us to use a spectral mapping theorem for the one-defect model. We also derive the time-averaged limit measure for one-dimensional case as an application of the spectral analysis., Comment: 19 pages

We describe a closed operator functional calculus in Banach modules over the group algebra L 1 ( R ) and illustrate its usefulness with a few applications. In particular, we deduce a spectral mapping theorem for operators in the functional calculus, which generalizes some of the known results. We also obtain an estimate for the spectrum of a perturbed differential operator in a certain class.

Speech-to-Singing (STS) conversion aims at converting one's reading speech into his/her singing vocal. The prior work was mainly focused on transforming the prosody of speech to singing, however, there exist prominent differences between the spectra of speech and singing, which need to be transformed as well. In this paper, we propose to make use of parallel multi-speaker speak-sing data to develop a speaker-independent spectral mapping model, which is conditioned on i-vector to generate target speaker/singer identity. The model is therefore called speaker conditioned spectral mapping model. The converted singing spectra are then used together with prosody features to synthesize the target singing. We investigate the effectiveness of i-vector based average model adaptation to model the differences between speech and singing spectra for a specific speaker. The proposed model does not require parallel speak-sing data from target speakers during training. The experimental results conducted on NUS-48E and NUS-HLT-SLS database indicate that the proposed approach significantly outperforms the baselines in terms of quality and similarity.

A mapping technique was used in the present study to explore the biological and imaging characteristics of invasive breast cancer and normal breast tissues in Raman examination data and construct a diagnostic model for breast cancer. Raman examination data reflect the biochemical or molecular characteristics of the target tissues. A total of 45 specimens from patients with breast cancer who underwent surgery and 25 adjacent normal breast tissue specimens were included in the present study. Using the specimens, a total of 53 sets of mapping data and 2,597 pieces of Raman spectral data were obtained. The collected spectra were corrected and fitted, the Raman spectra were analyzed by robust statistical methods, and a diagnostic model was constructed using the k-Nearest Neighbor (KNN) method. The KNN classification method was applied to analyze the characteristics of the mapping test application. The percentage of outliers in the mapping data for malignant and normal breast tissues was 12.7 and 6.6%, respectively. The percentage of outlier data in the conventional single-point detection data for malignant and normal breast tissues was 24.5 and 26.0%, respectively. Analysis using a t-test identified a significant difference in the number of outliers between mapping and single-point detection for malignant (t=-6.169; P

We discuss the description of eigenspace of a quantum walk model $U$ with an associating linear operator $T$ in abstract settings of quantum walk including the Szegedy walk on graphs. In particular, we provide the spectral mapping theorem of $U$ without the spectral decomposition of $T$. Arguments in this direction reveal the eigenspaces of $U$ characterized by the generalized kernels of linear operators given by $T$., Comment: 17 pages

Firstly, we establish spectral mapping theorems for normal eigenvalues (due to Browder) of a -semigroup and its generator. Secondly, we discuss relationships between normal eigenvalues of the compact monodromy operator and the generator of the evolution semigroup on associated with the -periodic evolutionary process on a Banach space X, where stands for the space of all -periodic continuous functions mapping to X.

In this paper, we give the fractional powers version of spectral mapping theorems for various essential spectra of non-negative linear operators and non-negative linear relations.

Ionic liquids (ILs) are promising solvents for the treatment and processing of cellulosic biopolymer materials. Despite the discovery of numerous ILs capable of dissolving varying amounts of cellulose, a detailed understanding of the dissolution process in ILs is still lacking. In this study, a new approach, Raman spectral mapping, is employed to investigate the kinetics of cellulose decrystallization and dissolution in a prototypical solvent IL, 1-ethyl-3-methylimidazolium acetate. The cellulose utilized is in the form of mercerized cotton yarns. By monitoring Raman bands characteristic of the crystalline domains, we investigate the degree and spatial extent of crystallinity in IL treated yarns. To elucidate the interplay of kinetic (viscosity) and thermodynamic (solvation) factors in the decrystallization process, the treatment time and temperature were systematically varied. A detailed analysis of the experimental results reveals three distinct temperature-dependent decrystallization regimes including (i) homogeneous decrystallization and cotton yarn collapse, (ii) fast-heterogeneous decrystallization of the collapsed yarn, and (iii) slow heterogeneous decrystallization. These results are discussed within the framework of the current understanding of cellulose and polymer dissolution.

Originally developed to classify multispectral and hyperspectral images, spectral mapping methods were used to classify Light Detection and Ranging (LiDAR) data to estimate the vertical structure of vegetation for Fuel Type (FT) mapping. Three spectral mapping methods generated spatially comprehensive FT maps for Cabaneros National Park (Spain): (1) Spectral Mixture Analysis (SMA), (2) Spectral Angle Mapper (SAM), and (3) Multiple Endmember Spectral Mixture Analysis (MESMA). The Vegetation Vertical Profiles (VVPs) describe the vertical distribution of the vegetation and are used to define each FT endmember in a LiDAR signature library. Two different approaches were used to define the endmembers, one based on the field data collected in 1998 and 1999 (Approach 1) and the other on exploring spatial patterns of the singular FT discriminating factors (Approach 2). The overall accuracy is higher for Approach 2 and with best results when considering a five-FT model rather than a seven-FT model. The agreement with field data of 44% for MESMA and SMA and 40% for SAM is higher than the 38% of the official Cabaneros National Park FTs map. The principal spatial patterns for the different FTs were well captured, demonstrating the value of this novel approach using spectral mapping methods applied to LiDAR data. The error sources included the time gap between field data and LiDAR acquisition, the steep topography in parts of the study site, and the low LiDAR point density among others.

In this paper, we consider the dense correspondence of volumetric images and propose a convolutional network-based descriptor learning framework using the functional map representation. Our main observation is that the correspondence-steered descriptor learning improves dense volumetric mapping compared with the hand-crafted descriptors. We present an unsupervised way to find the optimal network parameters by aligning volumetric probe functions and the enforcement of invertible coupled maps. The proposed framework takes the one-channel volume as input and outputs the multi-channel volumetric descriptors using the cascaded convolutional operators, which are faster than the conventional descriptor computations. We follow the deep functional map framework and represent the dense correspondence by the low-dimensional spectral mapping for the functional transfer and dense correspondence using the linear algebra. We demonstrate that by using the correspondence-steered deep descriptor learning, the quality of both the dense correspondence and attribute transfer are improved in the extensive experiments.

In this paper, we propose a progressive spectral mapping learning algorithm for throat microphone (TM) speech enhancement. Unlike previous full-band spectra mapping algorithms, this algorithm divides the spectra mapping from TM speech to Air-conducted (AC) speech into two tasks, one is the voice conversion task, and the other is the artificial bandwidth extension task. Long short-term memory recurrent neural network (LSTM-RNN) is further deployed as the mapping model. Objective evaluation results show that the TM speech quality is improved when compared with conventional full-band spectra mapping framework and DNN-based mapping model.

We introduce a noise-adaptive feed-forward deep neural network (DNN) for single-channel speech enhancement. The goal is to better exploit individual noise characteristics while training a spectral mapping DNN. To this end, we employ noise-dependent adaptation vectors, which are obtained based on the output of an auxiliary noise classification DNN, to adjust the weights and biases of the spectral mapping DNN. The parameters of the spectral mapping DNN, noise classification DNN and adaptation vectors are estimated jointly during the training stage. During the enhancement stage, we combine a classical unsupervised speech enhancement algorithm with the proposed DNN-based approach to further improve the enhanced speech quality. Experiments show that the proposed method provides better enhancement performance than the selected benchmark algorithms.

Band parameters can effectively be used as a rapid precursor processing step to enhance and augment the use of imaging spectrometer data by scientists that are not spectroscopy experts. Simple algorithms can be applied to large regional areas with consistent interpretation of surface composition and mineral exposures. Where these methods have been tested, they provide good agreement with results from more time consuming spectral feature fitting and detailed spectral mapping methods. We have tested band depths determined using three channels and derivative methods that have been shown to work well for CRISM.

We report the analysis of Hubble Space Telescope (HST) UV–optical spectroscopy of the nova-like variable V348 Puppis with eclipse mapping techniques. We measured the eclipse width at disc centre and determined a lower limit on the mass ratio of q > 0.3, with a tight relation between q and the binary inclination i. For 0.31 ≤ q ≤ 0.6, we have 79$_{.}^{\circ}$5 ≤ i ≤ 88$_{.}^{\circ}$2. Simulations with 3D eclipse mapping give no support to the suggestion of self-shielding of the accretion disc of V348 Pup, indicating a geometrically thin disc. Eclipse maps reveal two structures in the accretion disc interpreted as tidally induced spiral arms, which account for 50–60 per cent of the disc flux in the continuum. The uneclipsed component accounts for 30–50 per cent of the light, indicating that a significant fraction of the light comes from a vertically extended disc chromosphere + wind. Its spectrum can be fitted by an isothermal slab of gas with temperature T = 9600 K and column density Σ ∼1010 g cm−2, corresponding to optically thick thermal emission. Superhumps had faded by the epoch of the HST observations; the disc shrank in size while the disc wind flux increased with respect to observations when superhumps were present. This may be explained if the disc wind outflow is the dominant source of angular momentum loss from the accretion disc. Previously observed accretion disc radii of V348 Pup and of two other binaries with large discs are used to derive an empirical expression for the maximum possible disc radius, which implies that superhumps may be found in systems with up to q ≃ 0.7.

The spectrum and point spectrum of the Cesaro averaging operator \(\mathsf {C}\) acting on the Frechet space \(C^\infty ({\mathbb R}_+)\) of all \(C^\infty \) functions on the interval \([0,\infty )\) are determined. We employ an approach via \(C_0\)-semigroup theory for linear operators. A spectral mapping theorem for the resolvent of a closed operator acting on a locally convex space is established; it constitutes a useful tool needed to establish the main result. The dynamical behaviour of \(\mathsf {C}\) is also investigated.

We study truncated Toeplitz operators in model spaces K θ p for 1 p ∞ , with essentially bounded symbols in a class including the algebra C ( R ∞ ) + H ∞ + , as well as sums of analytic and anti-analytic functions satisfying a θ-separation condition, using their equivalence after extension to Toeplitz operators with 2 × 2 matrix symbols. We establish Fredholmness and invertibility criteria for truncated Toeplitz operators with θ-separated symbols and, in particular, we identify a class of operators for which semi-Fredholmness is equivalent to invertibility. For symbols in C ( R ∞ ) + H ∞ + , we extend to all p ∈ ( 1 , ∞ ) the spectral mapping theorem for the essential spectrum. Stronger results are obtained in the case of operators with rational symbols, or if the underlying model space is finite-dimensional.

We consider Ornstein–Uhlenbeck operators on L 2 ( R d ) perturbed by a radial potential V . Under weak assumptions on V we prove a spectral mapping theorem for the generated semigroup. The proof relies on a perturbative construction of the resolvent, based on angular separation, and the Gearhart–Pruss Theorem.

Understanding the relationship between the dynamics of neural processes and the anatomical substrate of the brain is a central question in neuroscience. On the one hand, modern neuroimaging technologies, such as diffusion tensor imaging, can be used to construct structural graphs representing the architecture of white matter streamlines linking cortical and subcortical structures. On the other hand, temporal patterns of neural activity can be used to construct functional graphs representing temporal correlations between brain regions. Although some studies provide evidence that whole-brain functional connectivity is shaped by the underlying anatomy, the observed relationship between function and structure is weak, and the rules by which anatomy constrains brain dynamics remain elusive. In this article, we introduce a methodology to map the functional connectivity of a subject at rest from his or her structural graph. Using our methodology, we are able to systematically account for the role of structural walks in the formation of functional correlations. Furthermore, in our empirical evaluations, we observe that the eigenmodes of the mapped functional connectivity are associated with activity patterns associated with different cognitive systems.

Visual and thermal cameras have large modality gap while providing different information of the same scene. The spectral mapping of thermal imagery into color imagery is a challenging task due to inherit nonlinear relationship. This paper deals with the automatic colorization of thermal imagery into color images using deep encoder-decoder convolutional neural network architecture. The presented approach is trained and evaluated on an online thermal/color dataset, where the network is trained on the spectral mapping of thermal to color images. The training is automatic without requiring user intervention. Extensive experimentation is carried out on the online available dataset, demonstrating consistent and harmonious appearance of generated colorized images from input thermal images.