Kyung-Woo Ko

Robust subpixel shift estimation using iterative phase correlation of a local region

In this paper, subpixel shift estimation method using phase correlation with local region is proposed for registration of noisy images. Commonly, phase correlation based on the Fourier shift property is used to estimate the shift between images. Subpixel shift of images can be estimated by the analysis for the phase correlation of downsampled images. However, in case of images with noise or aliasing artifacts, the error in estimation is increased. Thus, we consider a small region in a corner of an image instead of the whole, because flat regions with noise and regions with aliasing induce the error of estimation. In addition, to improve accuracy, the local regions are inversely shifted by varying the subpixel shift values, and obtaining the peak value of phase correlation between the images. Then, the subpixel shift value corresponding to the maximum of the peak values is selected. Real-time implementation of this process is possible because only a local region is used, thereby reducing the process time. In experiments, the proposed method is compared with conventional methods using several fitting functions, and it is applied for the task of super resolution imaging. The proposed method shows higher accuracy in registration than other methods, also, edge-sharpness in superresolved images is improved.

Hi-Fi Printer Characterization Method using Color Correlation for Gamut Extension

This paper proposes a colorimetric characterization method using the color correlation between the colorants in a hi-fi printer. While several colorant combinations can be used to match a certain color stimulus in a hi-fi printing system with more than 3 colorants, conventional colorimetric characterization methods only use 3 or 4 colorants to render a color, thereby limiting the color representation. As a result, the gamut is limited as they give up the other combinations of colorants. Therefore, this paper proposes a method of colorimetric characterization that uses combinations of all the colorants. As such, certain colorant combinations are selected based on considering the correlation factor between the colorant amount distributions. The correlation factor also affects the interpolation error, as the colorants are not independent of each other. Consequently, the total gamut is increased in low lightness regions, and the colors are represented more accurately.

Colorimetric characterization based on color correlation in CMYKGO printer

This paper proposes a method of colorimetric characterization based on the color correlation between the distributions of colorant amounts in a CMYKGO printer. In colorimetric characterization beyond three colorants, many color patches with different combinations of colorant amounts can be used to represent the same tri-stimulus value. Therefore, choosing the proper color patches corresponding each tri-stimulus value is important for a CMYKGO printer characterization process. As such, the proposed method estimates the CIELAB value for many color patches, then selects certain color patches while considering high fidelity and the extension of the gamut. The selection method is divided into two steps. First, color patches are selected based on their global correlation, i.e. their relation to seed patches on the gray axis, and become the reference for correlation. However, even though a selected color patch may have a similar overall distribution to the seed patch, if the correlation factor is smaller than the correlation factors for neighboring patches, the color patch needs to be reselected. Therefore, in the second step, the color patch is reselected based on the local correlation with color patches that have a lower correlation factor with the seed patch. Thus, to reselect the color patch, the seed patch is changed to the average distribution of eight neighboring selected color patches, and the new color patch selected considering the new correlation factor. Consequently, the selected color patches have a similar distribution to their neighboring color patches. The selected color patches are then measured for accuracy, and the relation between the digital value and the tristimulus value for the color patches stored in a lookup table. As a result of this characterization, the gamut is extended in the dark regions and the color difference reduced compared to conventional characterization methods.

Image mosaicking based on feature points using color-invariant values

In the field of computer vision, image mosaicking is achieved using image features, such as textures, colors, and shapes between corresponding images, or local descriptors representing neighborhoods of feature points extracted from corresponding images. However, image mosaicking based on feature points has attracted more recent attention due to the simplicity of the geometric transformation, regardless of distortion and differences in intensity generated by camera motion in consecutive images. Yet, since most feature-point matching algorithms extract feature points using gray values, identifying corresponding points becomes difficult in the case of changing illumination and images with a similar intensity. Accordingly, to solve these problems, this paper proposes a method of image mosaicking based on feature points using color information of images. Essentially, the digital values acquired from a real digital color camera are converted to values of a virtual camera with distinct narrow bands. Values based on the surface reflectance and invariant to the chromaticity of various illuminations are then derived from the virtual camera values and defined as color-invariant values invariant to changing illuminations. The validity of these color-invariant values is verified in a test using a Macbeth Color-Checker under simulated illuminations. The test also compares the proposed method using the color-invariant values with the conventional SIFT algorithm. The accuracy of the matching between the feature points extracted using the proposed method is increased, while image mosaicking using color information is also achieved.

Color recovery from gray image based on analysis of wavelet packet sub-bands

This paper proposes a colorization method that uses wavelet packet sub-bands to embed color components. The proposed method, firstly, involves a color-to-gray process, in which an input RGB image is converted into Y, Cb, and Cr images, and a wavelet packet transform applied to Y image to divide it into 16 sub-bands. The Cb and Cr images are then embedded into two sub-bands that include minimum information on the Y image. Once the inverse wavelet packet transform is carried out, a new gray image with texture is obtained, where the color information appears as texture patterns that are changed according to the Cb and Cr components. Secondly, a gray-to-color process is performed. The printed textured-gray image is scanned and divided into 16 sub-bands using a wavelet packet transform to extract the Cb and Cr components, and an inverse wavelet packet transform is used to reconstruct the Y image. At this time, the original information is lost in the color-to-gray process. Nonetheless, the details of the reconstructed Y image are almost the same as those in the original Y image because it uses sub-bands with minimum information to embed the Cb and Cr components. The RGB image is then reconstructed by combining the Y image with the Cb and Cr images. In addition, to recover color saturations more accurately, gray patches for compensating the characteristics of printers and scanners are used. As a result, the proposed method can improve both the boundary details and the color saturations in recovered color images.