This PDF file contains the editorial “Field testing” for OE Vol. 49 Issue 05

A new algorithm for no-reference blurred image quality assessment is proposed. This metric is based on block-based discrete cosine transform statistics and linear prediction method. We compare the performance of the proposed method with four blur metrics on three famous databases and our work shows the best results.

In Itti's model, which was one of the representative saliency models proposed in 1998, a Gaussian pyramid is used to analyze color information in scene images, and to generate a color conspicuity map. In this conspicuity map, some important objects can be located by salient areas, but their contours cannot be described clearly and perfectly. In this work, a wavelet low-pass pyramid is used to generate a color conspicuity map, and the contours of important objects pop out perfectly from salient areas. Experimental results validate the superiority of the proposed method.

We present a two-step program for point-by-point flat surface measurements with a fiber point diffraction interferometer (FPDI). The point diffraction wavefront reflected by a flat mirror under test is an aberrated spherical wavefront carrying the surface information of the flat mirror. The aberrated spherical wavefront interferes with a reference point diffraction wavefront through a plate beamsplitter (BS). The aberrations of the plate BS are also measured by the FPDI method. The figure of the flat mirror can be evaluated point by point after correcting the aberrations of the plate BS. This method makes use of the nearly perfect point diffraction wavefront, thus it can assure accurate flatness measurement on the optic under test. This research extends to the field of FPDI applications, and provides a new route for the high precision measurement of flat optics.

The alignment and collection of side-scattered light in scattering measurement on individual cells, based on hydraulic focusing technology, is always time-consuming and troublesome. Even small errors result in a poor SNR. This paper presents an optical system including a specially designed telecentric lens to focus stray light and scattering signals. The lens has a compact structure, small distortion and aberration, and large depth of focus. Thus scattering signals can be extracted easily and effectively, even when the spatial filters are out of focus or tilted with respect to the optical axis by human error. Also, the stray light in the side direction is analyzed, and the results show good performance of this special telecentric lens. Scattering results on 4.91- and 16.32-µm polystyrene beads verify that this optical system can be well applied for side-scattering signal detection.

We report initial results on designing and manufacturing a Schmidt-like corrector plate for a commercial off-the-shelf cylindrical lens, eliminating the cylindrical equivalent of its spherical aberration. The corrector is made by figuring the correction profile onto a precision glass window, which is subsequently aligned to the cylindrical lens. We have successfully fabricated the first plate and applied it in an interferometric test of a near-cylinder optic. The interferometric data from before and after applying the corrector demonstrates that the modified optic produces a cylindrical test wavefront with <1λ P-V of residual error at 632.8 nm, a >25× reduction compared to the uncorrected case.

We present a high-performance cover lens for light-emitting diode (LED) light sources to be applied to large-scale liquid crystal device television (LCD-TV) backlight modules. The cover lens with simple structure can redistribute light rays emitted from an LED light source to make the light distribution more extended so that the luminance uniformity of the backlight module can be increased. To gain a better performance, the parameters of the lens structure are optimized using a genetic algorithm. The simulation result shows that using the cover lens can achieve a luminance uniformity of 94.75% and an optical efficiency of 72.29% for a 46-in LCD-TV backlight module.

Parametric cost models are routinely used to plan missions, compare concepts, and justify technology investments. However, great care is required. There is a lot of confusion and wrong information for space telescopes. Cost estimating relationships based on primary mirror diameter vary by an order of magnitude. Cost estimating relationships based only on mass lack sufficient detail to support concept analysis and can lead to inaccurate conclusions by encouraging excessively complex and technologically immature solutions. Similarly, using ground-based models leads to incorrect conclusions. This work surveys current and historical published cost models for space telescopes while attempting to interpret them in a common logical framework to enable a systematic intercomparison.

The alignment of optical components is a key factor when designing and manufacturing multifiber optical systems. This problem can be treated as a standard multiobjective optimization problem and solved by numerical optimization methodologies. The core diameter of a single-mode fiber is ~9 µm, and any slight misalignment during manufacturing will cause signification optical losses in connections. Previous studies have shown that the currently used alignment methods for multifiber devices can increase the optical power summation of all fibers, but the results are not very accurate. This study first compares different numerical optimization methodologies that can be used to find the ideal connection position. Two indices are used to judge the performances of different methods: the required time and the optical power. Next, a two-stage optimization strategy is proposed to obtain a fast and accurate result. In the first stage, the Nelder-Mead simplex method is used to move toward the optimum position quickly. In the second stage, the steepest descent method with polynomial interpolation is applied to improve the accuracy because of the stability of the method.

A new, simple method to measure the effective focal length of a thick lens or lens system in the laboratory when a lens nodal bench is not readily available is described. Use is made of the Talbot autoimaging effect by placing the lens in a collimated light beam, with a Ronchi ruling in front of the lens and another in the refracted beam.

Channeled spectropolarimetry measures the complete polarization state of light, using a single spectrum, by amplitude modulating the Stokes parameters onto spectral carrier frequencies. However, spectral features that are not band limited in the Fourier transform domain, such as narrow-band atomic absorption lines, can generate false polarimetric signatures. We present a false-signature (aliasing) reduction technique that reduces the error induced by these non-band-limited features. Additionally, the spectral resolution of the S0 Stokes parameter is improved, up to the maximum resolution offered by the spectrometer. A theoretical model for implementing the technique on a Fourier transform infrared spectrometer is presented, including an enhanced model that accounts for dichroism within the crystal. The approach is experimentally demonstrated in the middle-wavelength infrared (3-5 µm) with the use of two multiple-order yttrium vanadate retarders. Additional results demonstrating the technique for wavelengths spanning 2.5-15 µm are obtained using cadmium sulfide retarders. Reconstructions are compared to conventional channeled spectropolarimetric reconstructions from the same system.

Two integral registration methods based on artificial features are described. In method one, independent global control points are designed to build a global coordinate system. Registration target and camera are also introduced to create intermediary coordinate systems. For each local scanning, one image of the whole measuring scene is shot by registration camera. Then local data can be unified to the global coordinate system by solving transition chains of various coordinate systems from this single image based on the projective geometry principle. In the other method, control points are placed on the object surface evenly and shot by registration camera from different positions and orientations. We solve their coordinates by employing the bundle adjustment method to build a global control network. The range sensor shoots at least three control points during each local scan. Then registration can be completed by mapping these control points into the global control network. In this work, the range sensor is untracked. Error accumulation and propagation are also effectively conquered, since overlapping of neighboring subregions is unessential. Experimental results are presented to show the feasibility of the proposed methods.

We present a technique for the focal length measurement of lenses applying Newton's method, using a cyclic path optical configuration (CPOC) and a wedge shear plate (WSP). Two point sources of orthogonal linearly polarized light waves that have a known longitudinal separation between them are generated using CPOC. Selecting a particular state of polarization (s or p), the corresponding point source is made to coincide with the back focal point of the test lens (TL). The collimation of the output beam from TL is tested using the WSP. With a change in the state of polarization, a spherical wavefront would emerge from TL because the corresponding point source is situated slightly away from the focal plane of TL. The radius of curvature of the emerging spherical wavefront from TL approximated as image distance is evaluated by analyzing the shear fringes produced by WSP. Appropriate formulae for the determination of the focal length of TL, from the image distance and known longitudinal separation between the sources, are derived using Newton's method. Results obtained for a positive doublet lens of nominal focal length of 500.0 mm is presented.

A new technique of wedge angle measurement of a transparent nearly parallel plate (PP) is presented. In this technique, the single-pass angular deviation suffered by a pair of mutually parallel, laterally separated pencil beams is enhanced by using a multipass cavity formed with two right-angle prisms. The resultant angular deflection is converted to linear deviation by means of a Fourier transform lens (FTL), which focuses the pencil beams on to its focal plane. A two-dimensional CCD detector array placed at the focal plane of the FTL measures the linear shift of the intensity centroid (IC) of the resulting Young's fringes formed due to the superposition of the laterally separated pencil beams. The wedge angle is determined from the measured value of the linear shift of the IC due to the PP.

The major problem of measurement of a power spectral density (PSD) distribution of surface heights with surface profilometers arises due to the unknown modulation transfer function (MTF) of the instruments, which tends to distort the PSD at higher spatial frequencies. The special mathematical properties of binary pseudo-random patterns make them an ideal basis for developing MTF calibration test surfaces. Two-dimensional binary pseudo-random arrays have been fabricated and used for the MTF calibration of the MicroMap^TM-570 interferometric microscope with all available objectives. An investigation into the effects of fabrication imperfections on the quality of the MTF calibration and a procedure for accounting for such imperfections are presented.

The dynamic angle limited integrated scattering (DALIS) method has been developed to examine optically smooth reflective surfaces with well-defined form. The DALIS system shows advantages over the conventional angle-resolved scattering. We propose a new configuration and results in the DALIS method by using a spherical mirror as a collecting element of the scattered light from the surface of a sample under test. Furthermore, the proposed method improves the detection of the scattered light and is suitable to be applied in workshop inspection during optical polishing processes.

We present simulations on temperature dependence of the threshold current density (J_th) in InGaN-AlGaN and ZnCdSe-ZnMgSSe quantum wire lasers having exciton transitions. We find that J_th in a quantum wire laser is insensitive to temperature variation when exciton binding energies are in the range of 30 to 50 meV. This behavior is similar to quantum dot lasers having free carrier transitions. We show that the temperature dependence of the threshold current density is greatly reduced in lasers having exciton transitions in comparison to the ones having free carrier transitions. The reduced dependence of J_th on the temperature is attributed to the inherent confinement of electron-hole pairs within the exciton radius along the wire axis, whereas the quantum wire cross section provides confinement of carriers in the plane perpendicular to the wire axis. The large exciton binding energies give rise to high exciton density of states and narrow spectral line width, thus confining the carriers into a pseudoquantum dot-like structure. The effect of exciton localization due to the randomness in quantum wire fabrication, causing exciton inhomogeneous line broadening, is also discussed.

We used the invariance of moving laser speckle patterns to develop a finger-controlled navigation system. Laser speckle images were captured using a commercially available image sensor. The relationships between template image size, available movement range, and a maximum correlation coefficient were derived. An essential criterion for setting the correlation coefficient threshold value was determined, and a displacement estimation algorithm was developed. A navigation system test platform and an accuracy analysis program were developed to monitor system performance related to variations in the template image size and the correlation threshold value. The experimental results suggest that the proposed design would be suitable to guide further development of finger-controlled navigation systems using laser speckle patterns.

We have proposed a novel biosensor scheme based on a dual photonic crystal microcavity with a stadium-shaped ring resonator (PCM-SSRR) and characterized the resonance conditions for the first time. The Q of the dual PCM-SSRR was significantly enhanced when a single PCM or a dual PCM was in the same resonance condition with the SSRR. We demonstrated a mutual interference between the dual PCM. In this study, the Q of the dual PCM-SSRR was increased three times over that of the directional coupler structure used with the single PCM. Assuming a detectable spectral resolution of 10 pm, a refractive index resolution of 3.03×10^-5 can be achieved on the dual PCM-SSRR.

In recent decades, optical fiber has proven useful for many sensor applications. Specifically, fiber Bragg grating (FBG) sensors have shown great utility for integrity management and environmental sensing of composite structures. One major drawback of FBG sensors, however, is the lack of a robust, nonpigtail technique for coupling to the embedded FBG sensor. In this paper, a novel method of free-space passive coupling of light into FBG sensors is described. An angled 45-deg mirror integrated directly into the fiber was used as an input coupling technique. We investigated the application of this approach to both single- and multimode glass fibers containing FBGs. For multimode FBGs, we studied the grating's uniformity across the fiber diameter and its effect on normal free-space coupling. In single-mode investigations, a novel method of coupling to the sensor via splicing a multimode fiber to a single-mode FBG (SMFBG) was developed. Finally, free-space coupling to an embedded SMFBG was employed to measure the tensile strain. Excellent agreement was found between the FBG and conventional electrical resistance strain gauges. We conclude that this coupling method might eliminate the need for pigtailing by providing a more robust coupling method for FBG sensors.

Surface plasmon resonance (SPR) sensors have been studied thoroughly for the past two decades. However, we found that the angular sensitivity in a prism-coupled SPR sensor can be as high as 500 deg/refractive index unit (RIU), which is two times higher than the sensitivity that has ever been achieved in previous studies. Such a high angular sensitivity can be fully achieved by simply choosing a proper low-index prism and a sufficiently large resonant angle for the light signal at an appropriate wavelength with an optimal metal film thickness. A feasible implementation of such an SPR sensor design concept was also proposed, and an even higher sensitivity of 600 deg/RIU can be achieved.

We have sought a suitable coaxial refractive index profile with dispersion-flattened and dispersion-shifted criteria in the context of a single-mode Raman gain amplifier. We present a novel comparative study based on the effective core area and Raman gain coefficient in relation to the effect of aspect ratios of trapezoidal index profiles in the inner core on the Raman gain.

We propose a novel system for generating multiplexed dark-soliton pulses using multiple light sources via an optical multiplexer, whereby dark solitons with different center wavelengths can be generated. The multiplexed signals can be transmitted into the communication link and filtered by using an optical add-drop filter. By using suitable simulation parameters, we have shown that dark solitons with different center wavelengths can in fact be obtained. In application, the communication capacity can be increased by using the multiplexed dark solitons; moreover, the transmission signals can be secured by utilizing the behavior of dark solitons. A free spectrum range (FSR) 2.5 nm and an amplified power of 30 W of dark solitons with wavelength 1.60 µm can be achieved. Channel spacing of the communication signals within a wavelength router can be provided by using a suitable FSR, which can be managed by using crosstalk effect analysis.

An algorithm based on phase spectrum analysis is proposed that can be used to correct the timing distortion between the multiple parallel demultiplexed post-sampling pulse trains in wavelength demultiplexing analog-to-digital converters. The algorithm is theoretically presented and its operational principle is explained. The algorithm is then applied to two parallel demultiplexed post-sampling signals from a proof-of-principle system and fairly good results are obtained. This algorithm is potentially applicable in other opto-electronic hybrid systems where an interleaving and/or multiplexing mechanism is utilized, such as optical time-division multiplexing and optical clock division systems, photonic arbitrary waveform generators, and so on.

The experimental work presented in this paper investigates the quality improvement of gain-switched (GS) sources. GS is a straightforward technique to achieve short pulses from diode lasers; however, such pulses exhibit long duration (10-100 ps), low power (in the milliwatt range), are asymmetric and are often generated along with pedestals or subpulses. Simultaneous compression and reshaping of these low quality pulses has been observed experimentally for several input power values. These effects have been achieved using a highly nonlinear optical-loop mirror, based on a microstructured optical fiber and a nonlinear semiconductor optical amplifier. It is designed to be compact and to directly process these complex pulses thus eliminating the requirement for prepulse conditioning. This offers overall benefits in terms of reduced system complexity. The quality of the pulses is characterized by their temporal width, pedestal height, duration, and spectral components. A maximum overall compression factor of 6 for a 10-mW average input power is observed. Additionally, the pulses obtained are observed to have improved quality for the input power range considered (500 µW to 30 mW, GS typical operational range) and the experimental scheme used preserves the compactness, efficiency, and simplicity of GS light sources.

We report on a high-speed time-domain optical coherence tomography (TD-OCT) technology that is capable of imaging 10 fps with 2000 A-scan. A duplex scanning optical delay line that is driven by a 1-kHz sinusoidal wave function was implemented and characterized. The reference and sample arms in a Michelson interferometer were simultaneously scanned using two identical piezoelectric transducers (PZTs). We adopted a sinusoidal wave function with constant voltages as the input functions for the PZTs. A recalibration of the measurement depth was necessary for the sinusoidal wave function operation. A recalibration with a spatial transformation was performed on the A-scan. We achieved an A-scan speed of 6 m/s. To the best our knowledge, this is the highest speed achieved in conventional TD-OCT systems to date.

The digital holographic microscope (DHM) has emerged as a useful tool for verifying the three-dimensional structure of an object. A dual-type inline DHM that can be used with both transmission and reflection imaging in a single device is developed. The proper modes (between transmission and reflection imaging) can be easily changed according to the characteristics of the object in this system. The optimum condition for retrieving the correct phase information is illuminating a plane wave to an object. In contrast to the transmission imaging, it is difficult to illuminate an object using a plane wave without deformations in the reflection imaging. We developed an adequate relay lens module for illumination that can be adapted to any type of microscope objective without significant aberrations in the reflection imaging. The relationship between the illuminating condition and the measured phase based on the wave optics is analyzed. A specially designed and manufactured phase mask is observed in this system, and an alternative method for overcoming the limitation of phase unwrapping is introduced for the inspection of that object.

A diffractive optical element with self-imaging capabilities is used to make a phase-shifting digital holography optical system simpler and cheaper. Sequential phase-shifting requires multiple exposures, and parallel phase-shifting demands a more complicated optical system. As opposed to typical phase-shifting methods, using the self-imaging diffractive optical element requires only one exposure on a low-cost CMOS matrix, and due to the small number of needed elements, the optical system is very compact. Instead of the approximation and interpolation methods, the properties of the self-imaging effect are utilized in the recording process and in the numerical reconstruction process.

For multilayer printed circuit board (PCB) and large-scale integrated circuit (LIC) chips, nondestructive testing of the inner structure and welding defects is very important for circuit diagram reverse design and manufacturing quality control. The traditional nondestructive testing of this kind of plate-like object is digital radiography (DR), which can provide only images with overlapped information, so it is difficult to get a full and accurate circuit image of every layer and the position of the defects using the DR method. At the same time, traditional computed tomography scanning methods are also unable to resolve this problem. A new reconstruction method is proposed for the nondestructive testing of plate-like objects. With this method, x rays irradiate the surface of the reconstructed object at an oblique angle, and a series of projection images are obtained while the object is rotating. Then, through a relevant preprocessing method on the projections and a special reconstructing algorithm, cross sections of the scanning region are finally obtained slice by slice. The experimental results prove that this method satisfactorily addresses the challenges of nondestructive testing of plate-like objects such as PCB or LIC.

A computationally efficient adaptive two-stage Karhunen-Loeve transform (KLT) scheme for spectral decorrelation in hyperspectral lossy bandwidth compression is presented. The component decorrelation of the JPEG 2000 (extension 2) is replaced with an adaptive two-stage KLT scheme. The data are partitioned into small subsets. The spectral correlation within each partition is removed via a first-stage KLT. The interpartition spectral correlation is removed using a second-stage KLT applied to the resulting top few sets of equilevel principal component (PC) images. Since only a fraction of each equilevel first-stage PC images are used in the second stage, the KLT transformation matrices will have smaller sizes, leading to further improvement in computational complexity and coding efficiency. The computation of the proposed approach is parametrically quantified. It is shown that reconstructed image quality, as measured via statistical and/or machine-based exploitation measures, is improved by using a smaller partition size in the first-stage KLT. A criterion based on the components of the eigenvectors of the cross-covariance matrix is established to select first-stage PC images, which are used in the second-stage KLT. The proposed scheme also reduces the overhead bits required to transmit the covariance information to the receiver in conjunction with the coding bitstream.

The line-fitting problem has been transposed to the signal-processing framework: Array-processing methods can be applied to virtual signals generated from the image, to estimate straight-line orientations. This paper deals with the estimation of straight and distorted lines in images by fast array-processing methods. Hough transform and snake methods retrieve straight lines and distorted contours, but present limitations. We adapt a fast high-resolution method, the propagator method, to the estimation of multiple distorted contours. For the first time, a method is proposed to cope with the intrinsically limited size of images, which reduces the accuracy of the high-resolution methods due to the low number of signal realizations. Moreover, an extension to images impaired by correlated noise is proposed. For this, an extension of the subspace-based methods to a method based on higher-order statistics is proposed. Distorted contours are assimilated to distorted wavefronts and retrieved with a novel optimization method. The performance of the proposed method is validated on several images.

In scene-based nonuniformity correction (NUC) methods for infrared focal-plane array cameras, the problem of ghosting artifacts widely affects the sensitivity of the imaging system and visibly decreases the image quality. Ghosting artifacts can also degrade the performance of several applications, such as target detection and tracking. We carried out a detailed analysis of the problem using a well-established NUC technique: the least mean square Scribner's algorithm. In order to solve some drawbacks of the original Scribner's algorithm, we introduced in the scheme a new technique that mitigates ghosting. Such technique relies on the employment of an edge-preserving spatial filter for the purpose of computing reliable spatial estimates. We tested the effectiveness of the new technique applying the improved NUC method to an experimental IR sequence of frames acquired in the laboratory. Finally, the performance of the proposed method was discussed and compared to that yielded by a well-established deghosting technique.

In order to improve signal-to-noise ratio (SNR) and contrast-to-noise ratio, we introduces a novel tunable forward-and-backward (TFAB) diffusion approach for image restoration and edge enhancement. In the TFAB algorithm, an alternative forward-and-backward (FAB) diffusion process is presented, where it is possible to better modulate all aspects of the diffusion behavior and it shows better algorithm behavior compared to the existing FAB diffusion approaches. In addition, there is no necessity to laboriously determine the value of the gradient threshold. We believe the TFAB diffusion to be an adaptive mechanism for image restoration and enhancement. Qualitative experiments, based on various general digital images and a magnetic resonance image, show significant improvements when the TFAB diffusion algorithm is used versus the existing anisotropic diffusion and the previous FAB diffusion algorithms for enhancing edge features and improving image contrast. Quantitative analyses, based on peak SNR and the universal image quality index, confirm the superiority of the proposed algorithm.

We discuss deinterlacing results in a computationally constrained and varied environment. The proposed computation-aware algorithm selection approach (CASA) for fast interlaced to progressive conversion algorithm consists of three methods: the line-averaging (LA) method for plain regions, the modified edge-based line-averaging (MELA) method for medium regions, and the proposed covariance-based adaptive deinterlacing (CAD) method for complex regions. The proposed CASA uses two criteria, mean-squared error (MSE) and CPU time, for assigning the method. We proposed a CAD method. The principle idea of CAD is based on the correspondence between the high and low-resolution covariances. We estimated the local covariance coefficients from an interlaced image using Wiener filtering theory and then used these optimal minimum MSE interpolation coefficients to obtain a deinterlaced image. The CAD method, though more robust than most known methods, was not found to be very fast compared to the others. To alleviate this issue, we proposed an adaptive selection approach using a fast deinterlacing algorithm rather than using only one CAD algorithm. The proposed hybrid approach of switching between the conventional schemes (LA and MELA) and our CAD was proposed to reduce the overall computational load. A reliable condition to be used for switching the schemes was presented after a wide set of initial training processes. The results of computer simulations showed that the proposed methods outperformed a number of methods presented in the literature.

A novel color-to-grayscale method is presented for converting color images to grayscale images by reducing dimensions. The proposed method converts three-dimensional (3-D) RGB color vectors into one-dimensional (1-D) grayscale values by projecting the 3-D vector into a two-dimensional (2-D) intermediate one followed by compressing the 2-D vector into the 1-D value. Characteristics of color are introduced to facilitate the final determination of the 1-D values in the reducing dimensions. The proposed method has the advantages of preserving chromatic contrasts, maintaining luminance consistency, and having a low computational cost. Furthermore, the proposed method has high resistance to artifacts, such as halos, which can occur when using local contents.

We propose a spectral imaging method for material classification and inspection of raw printed circuit boards (PCBs). The method is composed of two steps (1) estimation the PCB surface-spectral reflectances and (2) unsupervised classification of the reflectance data to make the inspection of PCB easy and efficient. First, we develop a spectral imaging system that captures high dynamic range images of a raw PCB with spatially and spectrally high resolutions in the region of visible wavelength. The surface-spectral reflectance is then estimated at every pixel point from multiple spectral images, based on the reflection characteristics of different materials. Second, the surface-spectral reflectance data are classified into several groups, according to the number of PCB materials. We develop an unsupervised classification algorithm incorporating both spectral information and spatial information, based on the Nystrom approximation of the normalized cut method. The initial seeds for the Nystrom procedure are effectively chosen using a guidance module based on the K-means algorithm. Low-dimensional spectral features are efficiently extracted from the original high-dimensional spectral reflectance data. The feasibility of the proposed method is examined in experiments using real PCBs in detail.

Three-dimensional (3-D) video systems are expected to be a next-generation visual application. Since multiview video for 3-D video systems is composed of color and associated depth information, its huge requirement for data storage and transmission is an important problem. We propose a rendering-oriented multiview video coding (MVC) method based on chrominance information reconstruction that incorporates the rendering technique into the MVC process. The proposed method discards certain chrominance information to reduce bitrates, and performs reasonable bitrate allocation between color and depth videos. At the decoder, a chrominance reconstruction algorithm is presented to achieve accurate reconstruction by warping the neighboring views and colorizing the luminance-only pixels. Experimental results show that the proposed method can save nearly 20% on bitrates against the results without discarding the chrominance information. Moreover, under a fixed bitrate budget, the proposed method can greatly improve the rendering quality.