Method of embedding and extracting watermark data using a correction image comparison with a noise-based image in the frequency domain
AssigneesBeyontech, Inc, Haroo Company
InventorJoon-ho Sung
Life of this patent
Estimating the term…
Abstract
Provided is a method of embedding and extracting watermark data that is robust against geometric distortion and low-quality photographing and for which the probability of successfully extracting watermark data for an original image is high, while the probability of successfully extracting the watermark data in the case of unauthorized copying is seriously impaired. The data embedding method according to an embodiment of the present invention comprises a step of converting the noise-based image using watermark data, and a step of adjusting the original image using the converted noise-based image.
Claims6
A data extracting method for extracting watermark data from a photographed image in which the watermark data is an embedded image, wherein the data extracting method comprises the steps of:
- generating a correction image from a captured image,
- generating a second two-dimensional result matrix,
- indicating whether a noise-based image is included in the photographed image comprising the steps of:
- performing a frequency domain process of comparing the photographed image with a noise-based image,
- transforming the second two-dimensional result matrix into a spatial domain to generate a second reference image,
- determining a reference point that is an extraction position of the watermark data in the captured image using the second reference image, and
- extracting the watermark data from the reference point of the photographed image with reference to the noise-based image;
- wherein, generating the second two-dimensional result matrix comprises the steps of:
- indicating whether the noise-based image is included in the corrected image by performing the frequency domain process of comparing the corrected image with the noise-based image;
- determining a reference point which is an extraction position of the watermark data comprising the steps of:
- determining the reference point in the corrected image;
- extracting the watermark data from a reference point of the corrected image;
- generating a correction image from the captured image comprising steps of performing a LogPolar transformation of the noise-based image and the captured image and calculating a scale and a rotation angle of the captured image relative to the noise-based image, and
- generating the corrected image from the photographed image by reflecting the calculated scale and the rotation angle.
A method, stored in a recording medium, for extracting watermark data from a captured image in which the watermark data are an embedded image, the method comprising the steps of:
- generating a correction image from a photographed image;
- monochrome-converting the correction image to maximize expression of noise inherent in the correction image and creating a monochrome-corrected image;
- generating a second two-dimensional result matrix indicating whether a noise-based image is comprised in the photographed image by performing a frequency domain process of comparing the photographed image with the noise-based image;
- transforming the second result matrix into a spatial domain to generate a second reference image;
- determining a reference point that is an extraction position of the watermark data in the captured image using the second reference image;
- and
- extracting the watermark data from a reference point of the photographed image with reference to the noise-based image;
- wherein, the step of generating a second two-dimensional result matrix comprises a step of generating a second result matrix indicating whether the noise-based image is comprised in the corrected image by performing a process in a frequency domain that compares the monochrome-corrected image with a noise-based image;
- wherein, the step of determining the extraction position of the watermark data comprises a step of determining a reference point that is an extraction position of the watermark data in the monochrome-corrected image;
- and
- wherein the step of extracting the watermark data comprises a step of extracting the watermark data from the monochrome-corrected image.
Dependencies are inferred from the claim text.