Image analysis – Color image processing
Reexamination Certificate
2001-03-12
2004-08-24
Tran, Phuoc (Department: 2621)
Image analysis
Color image processing
C382S165000, C382S168000
Reexamination Certificate
active
06782125
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a content-based multimedia retrieval method, and more particularly, to a content-based multimedia retrieval method, which retrieves multimedia data on the basis of color-based descriptors or texture-based descriptors.
2. Description of the Prior Art
Recently, active researches for a content-based multimedia retrieval are in progress.
In the content-based multimedia retrieval, a color-based descriptor such as a color histogram and a texture-based descriptor such as a texture histogram are mainly used as descriptors for retrieving the content-based multimedia data. The color-based descriptor is used as the most important information.
However, there is a disadvantage that the color-based descriptor may be easily distorted due to an illumination around image or characteristics of an image acquisition device. That is, although images are the same, if the images are obtained from different devices or the images illuminated by different lighting are obtained, people recognize the images as similar images, but an image analysis by a computer shows entirely different color information. It is called a distortion.
Because a multimedia data retrieval performance is deteriorated by the distortion, recently, researches for retrieving without an influence of the illumination are in progress.
In the meantime, conventional technology uses a retrieval method not considering color elements easily distorted by the illumination. Here, the color is expressed by several elements such as brightness, hue and chroma. Those are called color elements.
For example, in the conventional retrieval system, if a user wants to retrieve multimedia data regardless of brightness, the user retrieves data without using the brightness element out of the color elements (brightness, hue and chroma). If a user wants to retrieve multimedia data regardless of the illumination containing a specified hue, the user retrieves data without using the hue element.
However, the method for retrieving multimedia data regardless of the specified color elements (brightness, hue) can not help resulting in deterioration of an overall retrieval performance because it does not use given information and does not reflect the distortion of the color exactly.
As previously described, the conventional method has a problem that a performance of a retrieval system is generally deteriorated, because it retrieves multimedia data regardless of the illumination.
If there exists rules in the color distortion caused by the illumination or the characteristics of the device, it is possible to get a higher retrieval performance by analyzing and using the rules and taking a method for compensating the distortion suitable for the corresponding characteristics of each data.
Recently, researches for obtaining the optimum retrieval result by applying different weight to the descriptors used for content-based multimedia retrieval has been published.
The weight means an importance level of each descriptor or the elements of the descriptor. In retrieving multimedia data, higher weight is applied to the descriptor capable of more exactly retrieving.
The elements of the descriptors mean elements constituting the descriptor. For example, if the color histogram is the descriptor, histogram bins constituting the color histogram can be regarded as the elements of the descriptor.
The method for retrieving multimedia data by applying the high weight is expected to have a high retrieval performance theoretically, however, more effectively method for automatically obtaining the optimum weight has not yet been developed.
As described above, the conventional method for retrieving multimedia data regardless of a specified color element does not reflect color distortion information. Therefore, the conventional method has a problem resulting in deterioration of the overall retrieval performance.
SUMMARY OF THE INVENTION
Accordingly, the present invention is directed to a content-based multimedia retrieval method that substantially obviates one or more of the problems due to limitations and disadvantages of the related art.
To solve the above problems, an object of the present invention to provide a content-based multimedia data retrieval method, which is capable of increasing multimedia data retrieval performance by automatically extracting color distortion information of multimedia data and automatically setting weight to color-based descriptors on the basis of the extracted color distortion information.
It is another object of the present invention to provide a method for extracting color distortion information of multimedia data and a method for generating multimedia data containing the extracted color distortion information.
It is a further object of the present invention to provide a method for extracting a distortion sensitivity by differentially applying the weight according to the distortion sensitivity of a descriptor or elements of the descriptor.
To achieve the above objects, the present invention provides a content-based multimedia retrieval method, including the steps of: extracting color distortion information from multimedia data; setting weight of a descriptor used in the content-based multimedia retrieval using the extracted color distortion information; and retrieving multimedia data applying the weight.
To achieve the above object, in the content-based multimedia retrieval method, the weight is set as low as the color element has low chroma, considering a distortion sensitivity of the color element having low chroma is larger than that of the color element having high chroma. The weight is also set as low for the color element having the hue different from that of the color distortion information, considering the color region having the hue different from that of the color distortion information is largely distorted in hue.
To achieve another object, a color distortion information extracting method includes the steps of: dividing an image into partial regions of N×M; calculating, with respect to all divided partial regions, average hue values of pixels that their chroma is lower than a predetermined threshold for each partial region, and variance of hue values of pixels applied to the average hue values; and re-averaging the average hue values obtained in each partial region with respect to the partial regions in which the calculated variance is lower than the predetermined threshold.
To achieve a further object, in a content-based multimedia data retrieval using a descriptor such as a color histogram of image, a multimedia data information generating method, comprising the steps of: generating a descriptor group having one or more descriptors used in the multimedia data retrieval; and extracting color distortion information from the multimedia data, automatically setting weight of the descriptor group according to the extracted color distortion information, and generating color distortion information for retrieving the multimedia data.
To achieve a further object, a distortion sensitivity extracting method includes the steps of: calculating a similarity based on a descriptor or an element of the descriptor with respect to the same image data in a data sample group that a color distortion does not occur and in a data sample group that the color distortion occurs; obtaining an average value of the calculated similarities; and determining a distortion sensitivity in inverse proportion to the average similarity.
REFERENCES:
patent: 6373979 (2002-04-01), Wang
patent: 6564263 (2003-05-01), Bergman et al.
patent: 6643387 (2003-11-01), Sethuraman et al.
Kim Hyeon Jun
Lee Jin Soo
Fleshner & Kim LLP
LG Electronics Inc.
Tran Phuoc
LandOfFree
Content-based multimedia retrieval method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Content-based multimedia retrieval method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Content-based multimedia retrieval method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3272275