Yasunori Ishikawa

Optimization of Size of Pixel Blocks for Orthogonal Transform in Optical Watermarking Technique

We previously proposed a novel technology with which the images of real objects with no copyright protection could contain invisible digital watermarking, using spatially modulated illumination. In this “optical watermarking” technology, we used orthogonal transforms, such as a discrete cosine transform (DCT) or a Walsh-Hadamard transform (WHT), to produce watermarked images, where 1-b binary information was embedded into each pixel block. In this paper, we propose an optimal condition for a technique of robust optical watermarking that varies the size of pixel blocks by using a trade-off in the efficiency of embedded watermarking. We conducted experiments where 4 × 4, 8 × 8, and 16 × 16 pixels were used in one block. A detection accuracy of 100% was obtained by using a block with 16 × 16 pixels when embedded watermarking was extremely weak, although the accuracy did not necessarily reach 100% by using blocks with 4 × 4 or 8 × 8 pixels under the same embedding conditions. We also examined the effectiveness of using a Haar discrete wavelet transform (Haar DWT) as an orthogonal transform under the same experimental condition, and the results showed that the accuracy of detection was slightly inferior to DCT and WHT under very weak embedding conditions. The results from experiments revealed the effectiveness of our new proposal.

Tolerance Evaluation for Defocused Images to Optical Watermarking Technique

In this paper, we describe a new aspect to evaluating the robustness of the optical watermarking technique, which is a unique technology that can add watermarked information to object image data taken with digital cameras without any specific extra hardware architecture. However, since this technology uses light with embedded watermarked information, which is irradiated onto object images, the condition of taking a picture with digital cameras may affect the accuracy with which embedded watermarked data can be detected. Images taken with digital cameras are usually defocused, which occurs under non-optimal conditions. We evaluated the defocusing in images against the accuracy with which optical watermarking could be detected. Defocusing in images can be expressed with convolution with a line-spread function (LSF). We used the value of full-width at half-maximum (FWHM) of a Gaussian function as the degree to which images were defocused, which could approximate LSF. We carried out experiments where the accuracies of detection were evaluated as we varied the degree to which images were defocused. The results from the experiments revealed that optical watermarking technology was extremely robust against defocusing in images.

Optical watermarking technique robust to geometrical distortion in image

This paper presents an optical watermarking technique that is robust against geometrical distortion in images by using spatially modulated illumination. It can protect “analog” objects like pictures painted by artists from having photographs taken of them illegally in museums. Illegally captured images in practical situations may contain various distortions, and embedded watermarks may be incorrectly detected. Geometrical distortion caused by the shooting angle that the objects are captured at is a particularly major problem. We carried out experiments to evaluate the robustness of watermarking images that were geometrically distorted, in which distortions were intentionally created by moving the position of the projector and the digital camera from right in front of the object. The accuracy of the extracted watermarking data was almost 100%, even if the shooting angle was inclined by about 20 degrees between the projector and digital camera, in both cases when a Discrete Cosine Transform (DCT) and a Walsh-Hadamard Transform (WHT) were used as the methods of embedding watermarks. We introduced rectangular mesh fitting and a technique of “bi-linear interpolation” based on the four nearest points to correct the distortions.

Robust optical watermarking technique by optimizing the size of pixel blocks of orthogonal transform

We previously proposed a novel technology with which the images of real objects with no copyright protection could contain invisible digital watermarking, using spatially modulated illumination. In this “optical watermarking” technology we used orthogonal transforms such as a Discrete Cosine Transform (DCT) or a Walsh-Hadamard Transform (WHT) to produce watermarking images, where 1-bit binary information was embedded into each pixel block. Here, we propose a new robust technique of optical watermarking that varies the size of pixel blocks by a trade-off in the efficiency of embedded watermarking. We conducted experiments where 4×4, 8x8, and 16×16 pixels were used in one block. A detection accuracy of 100% was obtained by using a block with 16×16 pixels when embedded watermarking was extremely weak, although the accuracy did not reach 100% by using blocks with 4×4 or 8×8 pixels under the same embedding conditions. The results from experiments revealed the effectiveness of our proposed technique.

Optical watermarking robust to object with low-reflectance

We propose a new optically written watermarking technology that can prevent the illegal use of images of objects that cannot contain watermarking such as paintings. The new method involves DCT and iDCT of the whole area to increase robustness that enables the embedded information from watermarking to be accurately read out even when some parts of the information have been lost because of, for example, low reflectance from the objects surface. We conducted simulations and experiments where one-bit binary data were embedded in some tens of components in a high-frequency area. The experimental results revealed that embedded data could be read out extremely accurately at 100% even when up to 40% of the objects surface was dark. As a result, we demonstrated the feasibility of the technology we propose is practical.

Robustness against defocusing of images in optical watermarking technique

We describes in this paper the robustness of optical watermarking against the defocusing of images, which usually occurs in images taken with digital cameras under nonoptimal conditions. We evaluated measurements of the defocusing of images against the accuracy of detection of optical watermarking. The value of full width at half maximum (FWHM) of the Gaussian function was used to measure the defocusing of images. We found from the results of evaluation that optical watermarking technology was extremely robust against defocusing of images. As a result, we demonstrated the practicality of optical watermarking in a real-use environment along with the robustness against geometric distortion we previously proposed.

A technique of time domain sequential data embedding into real object image using spatially modulated illumination

We propose a new technique that uses spatially modulated illumination for embedding time domain sequential data into moving picture data. This is based on the “optical watermarking” technology that can be applied to protect “analog” objects like pictures painted by artists from having photographs taken of them illegally. Another important application of optical watermarking is embedding real time data sequence into real object images unconsciously, which may be moving pictures. We carried out experiments that sequential watermarking image data that had 10 fps (frame per second) were irradiated onto objects with projector, and moving image data with the picture rate of 30 fps were taken with a video camera. The result was that the embedded sequential watermarking information could be detected accurately from the obtained moving image data. In the experiments we embedded 256 bit information into watermarking image area of each frame that had the image size of 128 × 128 pixels, that is, the bit-rate of embedding information was 2.5kbps. The experimental results indicate that faster bit-rate and improved detection accuracy may be achieved when lager size of the area of embedding watermarks and higher resolution image for each frame are used.