Patent Application
20140037207
1. Field of the Invention
The present invention generally relates to an image sensor, and more particularly to a system and a method of adaptively suppressing false-color artifacts due to interpolation.
2. Description of Related Art
A Bayer color filter array (CFA), named after its inventor Dr. Bryce E. Bayer from Eastman Kodak, is widely used in digital image capture devices (such as digital still cameras), and is commonly placed over pixel sensors of a (black-and-white) image sensor to capture color image data. The Bayer CFA has a filter pattern typically made up of 50% green (G), 25% red (R) and 25% blue (B).
Although the Bayer CFA is capable of capturing color image data with low image capture cost, the Bayer CFA, however, imposes the need for interpolation of missing color data (i.e., the voids shown in
In order to suppress the false-color artifacts, an optical blur filter (or an anti-aliasing filter or an optical low-pass filter) is usually employed, as discussed in “Investigation of Color Aliasing of High Spatial Frequencies and Edges for Bayer-Pattern Sensors and Foveon X3® Direct Image Sensors” by Rudolph J. Guttosch, the disclosure of which is incorporated herein by reference.
Nevertheless, the blur filter reduces false-color artifacts at the expense of image sharpness. Further, use of the blur filter raises manufacturing cost. Moreover, inclusion of the blur filter in digital image capture devices increases device size and weight, making minimization in portable digital image capture devices, such as cell phones, less viable.
Accordingly, a need has thus arisen to propose a novel scheme of suppressing false-color artifacts in a more effective and economic manner.
In view of the foregoing, the embodiment of the present invention provides a system and a method of adaptively suppressing false-color artifacts due to interpolation on image data of an image sensor with a color filter array.
According to one embodiment, a system of adaptively suppressing false-color artifacts includes an interpolator, a converter, a luma spatial frequency unit, a chroma spatial frequency unit, a false-color detector and a compensator. The interpolator performs interpolation on raw image data outputted from an image sensor to obtain color interpolated data. The converter converts the interpolated data into a color space to result in converted data, each of which is represented by a luma component and at least one chroma component in the converted color space. The luma spatial frequency unit detects edges or an area with a substantially high luma spatial frequency within an active window, and the chroma spatial frequency unit detects substantial fluctuations in colors within the active window. The false-color detector determines occurrence of false-color artifacts according to results of the luma spatial frequency unit and the chroma spatial frequency unit. The compensator performs compensation on the chroma components if the false-color detector decides that the false-color artifacts occur, thereby resulting in compensated chroma components.
In the embodiment, in step 31, raw image data (e.g., red (R) data, green (G) data and blue (B) data) outputted from an image sensor (e.g., a Bayer image sensor) are interpolated by an interpolator 21, in order to obtain a set of complete red, green, and blue interpolated data R′/G′/B′ for each pixel.
Subsequently, in step 32, the interpolated data (e.g., R′/G′/B′) are then converted by a converter 22 into another color space, in which the converted data may be represented by a luma component and at least one chroma component. For example, in the embodiment, the interpolated data may be converted into YCbCr, in which Y represents a luma component, Cb represents a hue-difference chroma component, and Cr represents a red-difference chroma component. Although YCbCr is exemplified in the embodiment, it is appreciated that other color spaces, such as YUV, may be used instead. In YUV, Y represents a luma component, and UV represent two chroma components.
In step 33, the luma components (e.g., Y) and the chroma components (e.g., Cb and/or Cr) covered by an active window (or moving window) are fed to a luma spatial frequency unit 23A and a chroma spatial frequency unit 93B, respectively, for correspondingly obtaining a luma spatial frequency value (e.g., fY) of the luma components and at least one chroma spatial frequency value (e.g., fCb and/or fCr) of the chroma components. To be more specific, the luma spatial frequency unit 23A is utilized to determine a luma spatial frequency value fY of the luma components of the converted data covered, by an active window, for example, by a high pass filter (HPF), and the chroma spatial frequency unit 23B is utilized, to determine at least one chroma spatial frequency value fC of the chroma components of the converted data covered by the active window.
In step 34, the luma/chroma spatial frequency values (fY, fC) of the luma components and the chroma. components are then fed to a false-color detector 24 for determining occurrence (or presence) of false-color artifacts. In the embodiment, if (1) the luma spatial. frequency value fY of the luma components Y (within an active window) is greater than a predetermined luma threshold, indicating that a substantially high luma frequency is present, and (2) at least one chroma spatial frequency value fC of the chroma components of Cb and/or Cr (within the active window) is greater than a predetermined chroma threshold, indicating that a substantially high chroma frequency is present, the false-color detector 24 accordingly decides that false-color artifacts may occur. The rationale for using the false-color detector 24 to determine the occurrence of the false-color artifacts, in the embodiment, is based on the facts that the false-color artifacts normally occur in edges or high-spatial-frequency area (i.e., an area with a substantially high luma spatial frequency) and the false-color artifacts possess great fluctuations in colors (i.e., have a substantially high chroma spatial frequency).
If the false-color detector 24 decides that the false-color artifacts occur, the chroma components (Cb and/or Cr) are then subjected to compensation (step 35) by a compensator 25 in order to suppress false-color effect due to the interpolation, thereby resulting in compensated chroma components Cb′/Cr′. As a result, the fluctuations in colors due to the false-color artifacts may be substantially and effectively suppressed. On the other hand, if the false-color detector 24 decides that the false-color artifacts do not occur, the chroma components (Cb and Cr) are then passed without compensation. In other words, whether the chroma components (Cb and/or Cr) are subjected to compensation (step 35) by the compensator 25 is based on (or controlled by) a detection result of the false-color detector 24.
In one embodiment, the compensator 25 performs compensation in an amount (or an extent) in accordance with the luma spatial frequency value fY of the luma components. In another embodiment, the compensator 25 performs compensation in an amount in accordance with both (1) the luma spatial frequency value fY of the luma components Y and (2) at least one chroma spatial frequency value fC of the chroma components Cb and/or Cr.
Although specific embodiments have been illustrated and described, it will be appreciated by those skilled in the art that various modifications may be made without departing from the scope of the present invention, which is intended to be limited solely by the appended claims.
