Showing papers on "Upsampling published in 2005"

Parallel controllable texture synthesis

Abstract: We present a texture synthesis scheme based on neighborhood matching, with contributions in two areas: parallelism and control. Our scheme defines an infinite, deterministic, aperiodic texture, from which windows can be computed in real-time on a GPU. We attain high-quality synthesis using a new analysis structure called the Gaussian stack, together with a coordinate upsampling step and a subpass correction approach. Texture variation is achieved by multiresolution jittering of exemplar coordinates. Combined with the local support of parallel synthesis, the jitter enables intuitive user controls including multiscale randomness, spatial modulation over both exemplar and output, feature drag-and-drop, and periodicity constraints. We also introduce synthesis magnification, a fast method for amplifying coarse synthesis results to higher resolution.

Method and apparatus for supporting motion scalability

Abstract: A method and apparatus for supporting scalability for motion vectors in scalable video coding are provided. The motion estimation apparatus (120) includes a motion estimation module (121) searching for a variable block size and a motion vector that minimize a cost function for each layer according to predetermined pixel accuracy, a sampling module upsampling an original frame when the pixel accuracy is less than a pixel size, and before searching for a motion vector in a layer having a lower resolution than the original frame downsampling the original frame into the low resolution, a motion residual module calculating a residual between motion vectors found in the respective layers, and a rearrangement module rearranging the residuals between the found motion vectors and the found variable block size information using significance obtained from a searched lower layer. Accordingly, true motion scalability can be achieved to improve adaptability to changing network circumstances.

Enhanced resolution in SAR/ISAR imaging using iterative sidelobe apodization

Abstract: Resolution enhancement techniques in radar imaging have attracted considerable interest in recent years. In this work, we develop an iterative sidelobe apodization technique and investigate its applications to synthetic aperture radar (SAR) and inverse SAR (ISAR) image processing. A modified noninteger Nyquist spatially variant apodization (SVA) formulation is proposed, which is applicable to direct iterative image sidelobe apodization without using computationally intensive upsampling interpolation. A refined iterative sidelobe apodization procedure is then developed for image-resolution enhancement. Examples using this technique demonstrate enhanced image resolution in various applications, including airborne SAR imaging, image processing for three-dimensional interferometric ISAR imaging, and foliage-penetration ultrawideband SAR image processing.

Complete-to-overcomplete discrete wavelet transforms: theory and applications

Abstract: A new transform is proposed that derives the overcomplete discrete wavelet transform (ODWT) subbands from the critically sampled DWT subbands (complete representation). This complete-to-overcomplete DWT (CODWT) has certain advantages in comparison to the conventional approach that performs the inverse DWT to reconstruct the input signal, followed by the a/spl grave/-trous or the lowband shift algorithm. Specifically, the computation of the input signal is not required. As a result, the minimum number of downsampling operations is performed and the use of upsampling is avoided. The proposed CODWT computes the ODWT subbands by using a set of prediction-filter matrices and filtering-and-downsampling operators applied to the DWT. This formulation demonstrates a clear separation between the single-rate and multirate components of the transform. This can be especially significant when the CODWT is used in resource-constrained environments, such as resolution-scalable image and video codecs. To illustrate the applicability of the proposed transform in these emerging applications, a new scheme for the transform-calculation is proposed, and existing coding techniques that benefit from its usage are surveyed. The analysis of the proposed CODWT in terms of arithmetic complexity and delay reveals significant gains as compared with the conventional approach.

Reduced resolution update mode for advanced video coding

Abstract: There is provided a video encoder, video decoder and corresponding encoding and decoding methods for respectively encoding and decoding video signal data for an image slice. The video encoder includes a slice prediction residual downsampler (645) for downsampling a prediction residual of at least a portion of the image slice prior to transformation and quantization of the prediction residual. The video decoder includes a prediction residual upsampler (715) for upsampling a prediction residual of the image slice.

Scalable video coding method and apparatus using base-layer

Abstract: A method of more efficiently conducting temporal filtering in a scalable video codec by use of a base-layer is provided. The method of efficiently compressing frames at higher layers by use of a base-layer in a multilayer-based video coding method includes (a) generating a base-layer frame from an input original video sequence, having the same temporal position as a first higher layer frame, (b) upsampling the base-layer frame to have the resolution of a higher layer frame, and (c) removing redundancy of the first higher layer frame on a block basis by referencing a second higher layer frame having a different temporal position from the first higher layer frame and the upsampled base-layer frame.

Method and apparatus for encoding/decoding multi-layer video using DCT upsampling

Abstract: A method and apparatus for more efficiently upsampling a base layer to perform interlayer prediction during multi-layer video coding are provided. The method includes encoding and reconstructing a base layer frame, performing discrete cosine transform (DCT) upsampling on a second block of a predetermined size in the reconstructed frame corresponding to a first block in an enhancement layer frame, calculating a difference between the first block and a third block generated by the DCT upsampling, and encoding the difference.

Scalable Decoder And Expanded Layer Disappearance Hiding Method

Abstract: A scalable decoder which does not frequently switch the band of the decoded signal even if the signal in an expanded layer in band scalable encoding disappear and does not give any strangeness or discomfort to the subjective quality. If frame disappearance does not occur, the signal is a signal (S101). However, if a high-band packet is made to disappear, the actually received signal is only a low-band packet. Therefore, the scalable decoder subjects the signal of a low-band packet to an upsample processing. As a result, a signal (S102) where the sampling rate is a wide band and only the low-frequency component is left is generated. From the signal (S103) of the (n-1)-th frame, a compensation signal (S104) is generated by hiding and passed through an HPF to extract only the high-frequency component to generate a signal (S105). The signal (S101) where only the low-frequency component is left is added to the signal (S105) where high-frequency component is left to generate a decoded signal (S106).

Complete characterization of systems for simultaneous Lagrangian upsampling and fractional-sample delaying

Abstract: We present a complete formulation and an exact solution to the problem of designing systems for simultaneous sampling rate increase and fractional-sample delay in the Lagrangian sense. The problem may be regarded as that of a linear transformation, i.e., scaling, and/or shifting, of the uniform sampling grid of a discrete-time signal having a Newton series representation. It is proved that the solution forms a three-parameter family of maximally flat finite impulse response digital filters with a variable group-delay at the zero frequency. Various properties of the solution, including Nyquist properties and conditions for a linear phase response are analyzed. The solution, obtained in the closed form, is exact for polynomial inputs. We show that it is also suited for processing discrete-time versions of certain continuous-time bandlimited signals and signals having a rational Laplace transform. We then derive a generalization of the solution by augmenting the family with a fourth parameter that controls the number of multiple zeros at the roots of unity. This four-parameter family contains various types of maximally flat filters including those due to Herrmann and Baher. We list specific conditions on the four parameters to obtain many of the maximally flat filters reported in the literature. A significant part of the family of systems characterized by the solutions has been hitherto unknown. Examples are provided to elucidate this part as well.

Method and apparatus for performing sample rate conversion

Abstract: A digital modulator having a sample rate converter that does not require that the final output sampling rate, F OUT , of the sample rate converter be at least twice as great as the input sampling rate, F IN , of the sample rate converter. The upsampling interpolation filter stages of the sample rate converter of the invention do not require any reference clocks to be provided by the NCO of the sample rate converter. Therefore, the upsampling interpolation filter stages are decoupled from the final output sampling rate, F OUT . The interpolation algorithms may be implemented in software, hardware, or a combination of software and hardware. In addition, the polynomial-based interpolator of the Farrow Structure of the invention is capable of using an even or odd number of basepoints.

Method and apparatus for reconstructing the texture of a spatial enhancement-layer video picture

Abstract: Two texture prediction techniques used in Scalable Video Coding for reducing the redundancy between the spatial layers are upsampling of the reconstructed BL texture and upsampling of the BL residual. ESS aims at non-dyadic upsampling factors. A method for interpolating EL pixels (Q1, H, Q2) from BL pixels (E-2, E-1, E0, E1, E2, E3) comprises upsampling the BL elements (E-2, E-1, E0, E1, E2, E3) for usage as upsampled elements of the EL image, and if the BL image is a reconstructed image, then predicting according to the given upsampling factor sub-pel pixels (Q1, H, Q2) of the EL image by an interpolation based on n BL pixels (E-2, E-1, E0, E1, E2, E3), wherein different interpolation filters are used and wherein the respective filter coefficients approximate a windowed Lanczos-3 function.

Scalable picture encoding

Abstract: An encoding device (1) for encoding a picture signal (VS) comprises downsampling means (11) for converting the original picture signal into a downsampled picture signal, first encoding means (14) coupled to the downsampling means (11) for encoding the downsampled picture signal so as to provide a first encoded picture signal (BL), upsampling means (16) coupled to the downsampling means (11) for converting the downsampled picture signal into a upsampled picture signal, subtracting means (13) for producing a residual signal (RS) from the original picture signal and the upsampled picture signal, and second encoding means (14) coupled to the subtracting means (13) for encoding the downsampled picture signal so as to provide a second encoded picture signal (EL). The first encoding means (14) are arranged for lossy coding. The encoding device (1) is arranged for increasing the picture sharpness content of the residual signal (RS), and hence of the second encoded picture signal (EL), for example by using a filter (10).

Comparing interpolation schemes in dynamic receive ultrasound beamforming

Abstract: In medical ultrasound interpolation schemes are of- ten applied in receive focusing for reconstruction of image points. This paper investigates the performance of various interpolation scheme by means of ultrasound simulations of point scatterers in Field II. The investigation includes conventional B-mode imaging and synthetic aperture (SA) imaging using a 192-element, 7 MHz linear array transducer with λ pitch as simulation model. The evaluation consists primarily of calculations of the side lobe to main lobe ratio, SLMLR, and the noise power of the interpolation error. When using conventional B-mode imaging and linear interpolation, the difference in mean SLMLR is 6.2 dB. With polynomial interpolation the ratio is in the range 6.2 dB to 0.3 dB using 2nd to 5th order polynomials, and with FIR interpolation the ratio is in the range 5.8 dB to 0.1 dB depending on the filter design. The SNR is between 21 dB and 45 dB with the polynomial interpolation and between 37 dB and 43 dB with FIR filtering. In the synthetic aperture imaging modality the difference in mean SLMLR ranges from 14 dB to 33 dB and 6d B to 31 dB for the polynomial and FIR filtering schemes respectively. By using a proper interpolation scheme it is possible to reduce the sampling frequency and avoid a decrease in performance. When replacing linear interpolation with a more advanced interpolation scheme it is possible to obtain a reduction of 18 dB and 33 dB in the SLMLR for the B-mode and SA imaging, respectively, and an improvement in SNR of 24 dB. I. INTRODUCTION In medical ultrasound receive focusing is a core signal processing element used for reconstruction of image points from the received transducer signals in both conventional and synthetic aperture imaging. In the delay-and-sum beamformer a sample is selected from each of the receive channels corresponding to the echo of the image point target. The sample index is based on the total transmit-receive time-of- flight. Due to the continuous nature of the time-of-flight, it will not necessarily lie at the discrete time indices of the sampled channel data. Thus, some form of interpolation is needed and this heavily influences the image quality and the hardware complexity for implementation. By using a proper interpolation scheme it is possible to reduce the sampling frequency or to improve performance. The investigation in this paper is based on the work of Henrik Andresen (1) and quantifies the change in performance as a function of interpolation type applied by means of ultrasound simulations of point scatterers in Field II (2). This paper introduces the beamformation toolbox, BFT2 which is used in the investigation. BFT2 (3),(4), developed at CFU is written in C and has a Matlab program interface. It performs dynamic receive focusing and offers choices between static or dynamic apodization and various interpolation schemes. The interpolation schemes investigated include linear, polynomial, and upsampling and FIR filtering. Various order polynomials and FIR filters are investigated. II. DESCRIPTION The investigation in this paper on the influence of the choice of interpolation scheme includes conventional B-mode imaging and synthetic aperture imaging (SAI). The ultrasound RF signals for the investigation is created using Field II and the beamforming is performed with BFT2. From a reference data set an evaluation data set is created, which is used for the evaluation. For the evaluation the point-spread function, PSF is useful when observing the characteristics of different imaging modalities. It is highly affected by the type of transmit-receive focusing used, and is, thus, also affected by the interpolation in receive beamforming. The lateral part of the PSF is especially interesting in terms of spatial distribution and amplitude of the side-lobe energy, which again directly affects the image contrast. The evaluation and comparison of interpolation schemes is done by partly observing the lateral PSF and quantizing the main-lobe and side-lobe energy distribution in terms of the full-width-half-maximum, FWHM and the side-lobe-main-lobe-ratio, SLMLR and partly by the noise power of the interpolation error. The FWHM and the SLMLR are calculated in the horizontal plane at the depth of each of the point scatterers and compared to the case where the reference data is used. A. Simulation setup The ultrasound RF signals for the investigation is created us- ing Field II with a 192-element, 7 MHz linear array transducer with λ pitch as simulation model and a 3-cycle sinusoid as excitation. The simulation is based on point scatterers placed at a depth of 10 mm to 80 mm with 5 mm between each, placed along the center of the transducer. A reference RF data set has been created using a sampling frequency of 1 GHz and linear interpolation and the evaluation RF data set is created by decimating the reference data (picking out samples) to a sampling frequency of 40 MHz. Two data sets are created. One by using conventional B-mode imaging, and one by using

Parallel texture synthesis by upsampling pixel coordinates

Abstract: A method and system for synthesizing texture using upsampled pixel coordinates and a multi-resolution approach. The parallel texture synthesis technique, while based on a neighborhood matching technique having order-independent texture synthesis, extends that approach in at least two areas, including efficient parallel synthesis and intuitive user control. Pixel coordinates are upsampled instead of pixel colors, thereby reducing computational complexity and expense. These upsampled pixel coordinates then are jittered to provide texture variation. The jitter is controllable, such that a user has control over several aspects of the jitter. In addition, each neighborhood-matching pass is split into several sub-passes to improve correction. Using sub-passes improves correction speed and quality. The parallel texture synthesis system and method disclosed herein is designed for implementation on a parallel processor, such as a graphics processing unit.

Video decoding method using smoothing filter and video decoder therefor

Abstract: A method and apparatus for increasing output picture quality at a decoding terminal by using a smoothing filter are provided. The video decoding method includes generating a residual frame from an input bitstream, performing wavelet-based upsampling on the residual frame, performing non-wavelet-based downsampling on the upsampled frame, and performing inverse temporal filtering on the downsampled frame.

Method and apparatus for reconstructing a video frame for spatial multi-layer video sequence

Abstract: A method for reconstructing a video frame in a spatially scalable environment with a base-layer (BL) and an enhancement-layer (EL), wherein frames are intra-coded (I), predicted (P) or bi-directionally predicted (B), and wherein groups of frames (GOP) are defined containing per layer at least two low-pass frames (I,P), comprises reconstructing a missing or unusable frame by copying, interpolating, predicting or upsampling, wherein copying other frames or interpolating between other frames is used for reconstruction of a hierarchical B-frame, upsampling the motion information (EC4, EC10) and/or residual of a BL frame is used if the missing or unusable frame is an EL frame and its corresponding BL frame is available, and the corresponding BL frame is reconstructed and upsampled if the missing or unusable frame is an EL frame and both BL low-pass frames and not both EL low-pass frames of the current group are available.

Supergridded cone-beam reconstruction and its application to point-spread function calculation

Abstract: In cone-beam computed tomography (CBCT), the volumetric reconstruction may in principle assume an arbitrarily fine grid. The supergridded cone-beam reconstruction refers to reconstructing the object domain or a subvolume thereof with a grid that is finer than the proper computed tomography sampling grid (as determined by gantry geometry and detector discreteness). This technique can naturally reduce the voxelization effect, thereby retaining more details for object reproduction. The grid refinement is usually limited to two or three refinement levels because the detail pursuit is eventually limited by the detector discreteness. The volume reconstruction is usually targeted to a local volume of interest due to the cubic growth in a three-dimensional (3D) array size. As an application, we used this technique for 3D point-spread function (PSF) measurement of a CBCT system by reconstructing edge spread profiles in a refined grid. Through an experiment with a Teflon ball on a CBCT system, we demonstrated the supergridded volume reconstruction (based on a Feldcamp algorithm) and the CBCT PSF measurement (based on an edge-blurring technique). In comparison with a postreconstruction image refinement technique (upsampling and interpolation), the supergridded reconstruction could produce better PSFs (in terms of a smaller FWHM and PSF fitting error).

On Cooley-Tukey FFT method for zero padded signals

Abstract: The classical Cooley-Tukey fast Fourier transform (FFT) algorithm has the computational cost of O(Nlog2N) where N is the length of the discrete signal. Spectrum resolution is improved through padding zeros at the tail of the discrete signal, if (p -1)N zeros are padded (where p is an integer) at the tail of the data sequence, the computational cost through FFT becomes O(pNlog2pN). This paper proposes an alternate instance of padding zeros to the data sequence that results in computational cost reduction to O(pNlog2 N). It has been noted that this modification can be used to achieve non-uniform upsampling that would zoom-in or zoom-out a particular frequency band, in addition, it may be used for pruning the spectrum, which would reduce resolution of an unimportant frequency band

Sample rate conversion module and applications thereof

Abstract: A sample rate converter includes an upsampling module, a low pass filter, and a linear sample rate conversion module. The upsampling module is operably coupled up-sample a digital input signal having a first rate to produce a digitally up-sampled signal. The low pass filter is operably coupled to low pass filter the digitally up-sampled signal to produce a digitally filtered signal at an up-sampled rate. The linear sample rate conversion module is operably coupled to convert the digitally up-sampled signal into a sample rate adjusted digital signal having a second rate based on an control feedback signal and a linear function, wherein a relationship between the first rate and the second rate is a non-power of two.

Comparative study of interpolation techniques for ultra-tight integration of GPS/INS/PL sensors

Abstract: Ultra-tight architecture plays a key role in improving the robustness of the integrated GPS/INS/PL (Pseudolite) system by aiding GPS receiver's carrier tracking loops with the Doppler information derived from INS (Inertial Navigation System) velocity measurements. This results in a lower carrier tracking loop bandwidth and subsequent improvement in measurement accuracy. Some other benefits using this architecture include: robust cycle-slip detection and correction, improved anti- jam performance, and weak signal detection. Typically the integration/navigation filter run at a rate of 1 to 100 Hz, which is insufficient to aid the carrier tracking loop as such loops normally run at about 1000 Hz. Two approaches were envisioned to solve this problem. One approach is to run the navigation Kalman filter at a higher rate, and the other is to run the filter at a lower rate and interpolate the measurements to the required rate. Although the first approach seems to be straightforward, it is computationally very intensive and requires a huge amount of processing power, adding to the cost and complexity of the system. The second method interpolates the low rate Doppler measurements from the navigation filter using multirate signal processing algorithms. Due to its efficiency and simpler architectures the interpolation method is adopted here. Filtering is the key issue when designing interpolators as they remove the images caused in the upsampling process. Although direct form of filtering can be adopted, they increase the computations. To reduce the computational burden, two efficient ways of implementing the interpolators are proposed in this paper: Polyphase and CIC (Cascaded Integrator Comb). The paper summarizes the design and analysis of these two techniques, and our initial results suggest that CIC is relatively better in terms of performance and computational requirements.

A novel general formulation of up/downsampling commutativity

Abstract: The paper establishes the commutativity conditions of upsampling and downsampling for multidimensional signals defined on discrete Abelian groups (lattices). The general condition includes results available in the literature as particular cases. Complete generality is achieved by giving an abstract definition of up/downsampling and by working in the signal domain instead of the traditional approach based on z and Fourier transforms. Examples of applications are outlined 1) for the one-dimensional (1-D) case verifying existing results, 2) for television scanning, 3) for degenerate (reduced-dimensionality) lattices, and 4) for multiplicative groups.

Reducing aliasing in wavelets based downsampling for improved resolution scalability

Abstract: In wavelets based coding applications, resolution scalability is achieved by retaining the low pass signal subband corresponds to the required resolution and discarding other high pass wavelet subbands. Aliasing is a common problem present in such downsampling. In this paper a novel technique for improving the low pass filter for improved downsampling is presented. This method uses an extra update step followed by P+U lifting scheme. The preprocessing update step is chosen as the dual update step associated with the wavelet. The spatially adaptive low pass (SALP) filtering concept is used for the second update step, leading to an overall low pass filter whose size adapts to the underlying signal content. The filter choices for the second update step is recovered at the decoder without any bookkeeping. Results using the 2D 5/3 wavelet with the extra pre-processing update step show improvements over conventional wavelets.

Oversampled phased array beamformer

Abstract: A phased array beamformer is provided, comprising a plurality of sensors arranged in an array, a plurality of samplers for generating respective oversampled signals, a digital filter for performing delay and sum filtering of the oversampled signals to generate a beamformed output signal, and a decimator for downsampling the beamformed output signal to generate a Nyquist sample rate signal at a predetermined look direction. The delay and sum beamforming operation is performed at the oversampling frequency, preferably within a ΔΣ modulator scheme.

Efficient implementation of a spectrum analyzer for fixed point architectures

Abstract: This paper presents a novel and efficient algorithm to compute fractional band spectrum digital decomposition without using fast Fourier transformation (FFT). The algorithm is optimized for fixed point architectures in order to reduce the required computational power. It is based on lattice coupled all-pass IIR digital filters, downsampling and recursive techniques. Hence the proposed algorithm is capable of computing up to third band spectrum analysis on low cost processors. Though fixed-point arithmetic is used the spectrum analysis may be performed at very low frequency, irrespective of the sampling frequency, and very narrow band pass filters can be realized. The latter features are critical to achieve with traditional IIR structures. The proposed algorithm is patent pending.

Residue sampling for image and video compression

Abstract: The chrominance subsampling has been widely used for the image and video compression. However, this can cause color distortion and image quality degradation. We developed a new sampling method in which the residual image after intra or inter prediction is subsampled instead of the original image subsampling. Since the residual image contains fewer information than the original image does generally, the information loss is reduced when the residual image is downsampled. Also the upsampling procedure is much less sensitive to the filtering methods selected as demonstrated in the experimental results. So we can apply a simple filtering to reduce the complexity. The experimental results show the effectiveness of the proposed sampling method. This sampling method can be further improved by applying adaptively according to the statistics of each block, and can be applied to the RGB image directly without color conversion.

Image processing apparatus for downsampling and interpolating an image signal from a colour filter array sensor

Abstract: of EP1081937An image processing apparatus includes an interpolation unit, compensation value calculation unit, and compensation unit. For an interpolation point selected from an image signal at an interval of n pixels (n is an integer of 2 or more), the interpolation unit interpolates the pixel value of each color data at the interpolation point using the pixel values of pixels of the same color falling within an interpolation region of m x m pixels (m is an integer of 2 or more) centered on the interpolation point, and outputs the pixel value as an interpolated pixel value at the interpolation point for each color data. The compensation value calculation unit generates a pixel compensation value for compensating for the pixel value of the interpolation point using the pixel values of pixels falling within a compensation region of M x M pixels (M is an integer equal to or larger than m and n) centered on the interpolation point. The compensation unit compensates for the interpolated pixel value of each color data at the interpolation point that is output from the interpolation unit by using the pixel compensation value corresponding to the interpolation point that is obtained by the compensation value calculation unit, and outputs the compensated pixel value as a new pixel value of each color data at the interpolation point.

Method and Device for Coding a Sequence of Video Images

Abstract: A video image sequence is coded or decoded. By motion compensated temporal filtering, using discrete wavelet decomposition, the discrete wavelet is decomposed by dividing the video image sequence into source and destination groups of images. An image in the destination group is determined from at least one image including pixels in the first group of the source group. The representative image includes pixels and subpixels determined from pixels and subpixels obtained by upsampling at least one image in the source group.

Time domain upsampling linear MMSE receiver for asynchronous MC-CDMA systems over frequency-selective Rayleigh fading channels

Abstract: In this paper, a novel time domain upsampling linear receiver is proposed for an asynchronous multicarrier code division multiple access system in frequency-selective Rayleigh fading channels. After anti-aliasing filtering, the received discrete-time upsampled signal is processed by a finite-impulse-response (FIR) filter at the receiver without any pre-processing. For the FIR filter, the observation window is extended to two-symbol duration in order to capture a complete desired symbol and eliminate the knowledge of signature waveform and timing acquisition requirement, and the tap coefficients are optimised according to the minimum mean square error (MMSE) criterion by employing a normalised least-mean-square algorithm. The proposed time domain upsampling MMSE (TDU-MMSE) receiver works in a decision-directed mode after training by a pre-defined sequence. The theoretical bit error rate (BER) is derived for the proposed receiver and agrees very well with the simulation results. Moreover, it has been shown that the TDU-MMSE receiver is able to achieve better BER performance than the conventional linear MMSE (LMMSE) and partial sampling MMSE (PS-MMSE) with lower computation complexity.

Efficient Image Upsampling using Frequency Resolution Expansion Schemes in DCT Domain

Abstract: Image upsampling can be performed in both spatial and frequency (transform) domain. In the spatial domain, various upsampling techniques are developed and 6-tap FIR interpolation filter is most well known method, which is embedded in many video coding standards. It can provide high subjective quality but shows low objective quality. In the transform domain, simple zero padding method can produce upsampled image easily. It shows better objective quality than 6-tap filtering, but it yields ringing effects which annoy eyes. In this paper, we present efficient upsampling method using frequency addition method in transform domain to provide better subjective and objective quality than conventional method Extensive simulation results show that the proposed algorithm produces visually fine images with high PSNR.