CN102905127A - A video watermark implementation method - Google Patents

Abstract

The invention discloses a video watermark implementation method by which a video watermark is invisible during playing of a video and is invisible during pausing of the video. Watermark values which are of the same bit and have opposite signs are embedded into the same airspace positions of two adjacent frames of images in video with the frame rate being more than 60Hz according to an eye vision persistance characteristic and the characteristic that the similarity of two adjacent frames of images in video is high, thereby realizing the effects that the watermark is clear during pausing of the video, the watermark is invisible during playing of the video, and the subjective quality of the video is not influenced. By the method, the watermark is still effective and clear after compression according to an H.264/advanced video coding (AVC) compression encoding standard. An extraction algorithm and extraction equipment are not needed; and the video watermark implementation method has the advantages of low overhead, simplicity in use and effectiveness.

Description

A video watermark implementation method

technical field

The invention relates to the field of video information processing, and mainly relates to a method for realizing a visible video watermark when the video is still.

Background technique

Video watermarking is a watermark loaded on digital video, and it is an effective digital product copyright protection and data security maintenance technology. Video watermarking can be divided into perceptible and imperceptible. Perceivable digital watermarks are generally light or translucent unsightly patterns, which are mainly used to declare the ownership, copyright and source of the product on the spot, and play a role in advertising or restraint. The imperceptible digital watermark is like invisible text in invisible ink technology, hidden in digital products. At present, in the aspect of video watermarking research, most of the researches are non-perceived watermarks, and the related researches combining perceptible watermarks and non-perceived watermarks have not yet appeared.

Contents of the invention

In order to solve the above technical problems, the present invention provides a video watermark implementation method that is invisible when the video is played and visible when the video is still. The method has the advantages of simplicity and effectiveness.

In order to achieve the above object, the present invention adopts the following technical solutions.

Embed the same watermark value with the opposite sign in the same spatial position of two adjacent frames, and the applicable requirement for video is that the frame rate reaches 60Hz and above. This method divides the video frame into odd frames and even frames, adds (or subtracts) the binary watermark to the pixel values of the odd frames and subtracts (or adds) the pixel values of the even frames according to a certain ratio, so that adjacent The average value of the pixels at the same position in the two frames is equal to the pixel value of the original video image, achieving the complementary effect of the image pixel values at the watermark embedding place of the odd and even frames.

The invention includes three basic schemes: the first scheme is to embed the watermark into the luminance (Y) component of the video; the second and third schemes are to embed the watermark into the two chrominance (U, V) components of the video respectively . At the same time, other schemes can be derived from the combination of these three basic schemes.

The present invention embeds the watermark directly into the pixels in the image space. The watermark embedding method: before the watermark is embedded, the watermark value b is adjusted according to the pixel value of the carrier image to be embedded, and the positive and negative values of the watermark value b are determined according to the parity of the video frame, such as Formula 2: After the watermark value is determined, the embedding method of the watermark is as in formula 1. When the watermark information bit is 0, the pixel value of the carrier image is subtracted from the adjusted watermark value b; when the watermark information bit is 1, the pixel value of the carrier image is Added to the adjusted watermark value b.

$\stackrel{\‾}{x}=\left\{\begin{array}{c}x(i,j)-b,W=0\\ x(i,j)+b,W=1\end{array}\right.$ Formula 1

公式2

Formula 2

为水印嵌入后的像素值，W表示水印信息位。参数b表示要嵌入的水印值，符号根据要嵌入的视频帧的奇偶性进行调节，大小由像素值决定。α为权值因子，其值根据所选方案和视频自行设定。round（）函数表示输出该函数输入值(α*X(i，j))四舍五入后的整数值。视频帧为奇数帧时，b值为正整数+round（）；视频帧为偶数帧时，b值为负整数-round（）。Among them, X (i, j) is the pixel value of the video image,

is the pixel value after watermark embedding, and W represents the watermark information bit. The parameter b represents the watermark value to be embedded, the symbol is adjusted according to the parity of the video frame to be embedded, and the size is determined by the pixel value. α is a weight factor, and its value is set according to the selected scheme and video. The round() function indicates to output the rounded integer value of the function input value (α*X(i, j)). When the video frame is an odd frame, the value of b is a positive integer + round(); when the video frame is an even frame, the value of b is a negative integer - round().

Solution 1: Embed the watermark into the Y component of the video frame. The human eye is sensitive to the luminance component, and the quantization step size in the compression coding process is small, the watermark value embedded in the Y component is not easy to be too large, and the weight factor α value is less than 1/8.

Scheme 2: Embed the watermark into the U component of the video frame. The human eye is not sensitive enough to the chrominance component, the quantization step size in compression coding is large, the watermark value embedded in the U component should not be too small, and the weight factor α is less than 1/6 to meet the requirements.

Scheme 3: Embedding the watermark into the V component of the video frame. The third scheme is similar to the second scheme, and the selection of the weight factor α can refer to the second scheme. But the experiment shows that the change of V component is more noticeable than that of U component, so the watermark is more suitable to be embedded in the area or video frame with smaller pixel value of V component.

Specific method steps:

1) Read the binary watermark image;

2) Select the component to embed the watermark into the video frame, obtain the video frames of the video file one by one, and judge the parity of the video frame; determine the area to be embedded according to the size of the watermark image;

3) Adjust the sign of the watermark value according to the parity of the video frame, so that the sign of the same watermark value embedded in the same position of two adjacent frames is opposite, and embed the adjusted watermark value into the pixel of the video frame component;

4) Take the next video frame as the current video frame, and repeat step 3 until all video frames of the video file are processed.

In the step 2), select the component to embed the watermark into the video frame, and the selection of the component can be one or a combination of the following schemes:

a. Embed the watermark into the Y component of the video frame, and the weight factor α is less than 1/8;

b. Embed the watermark into the U component of the video frame, and the weight factor α is less than 1/6;

c. Embed the watermark into the V component of the video frame, and the weight factor α is less than 1/6.

The present invention utilizes the persistence of human vision and the high similarity between two adjacent frames of video to embed the watermark directly into the airspace of the video image, so that the watermark is clearly visible when the video is still and invisible when the video is playing. And does not affect the effect of the subjective quality of the video. At the same time, the present invention can still be effective and visible after being compressed by the H.264/AVC compression coding standard by adjusting the value of the weight factor α.

Beneficial effects of the present invention: the present invention proposes a brand-new idea and technical platform for video watermarking, which has the advantages of simple, effective, and high real-time embedding methods, and at the same time, the present invention has the characteristics and Advantages: The watermark is invisible when the video is playing; the watermark can be seen intuitively without other operations and equipment when it is still; the watermark is robust and clear, can resist noise and compression processing, and is still clearly visible after being attacked. Therefore, the present invention does not need a watermark extraction algorithm and an extraction device, and the extra cost is small. In addition, the present invention, as a brand-new technical platform, has good scalability, and technicians and developers can improve and modify it according to their own needs on the basis of the technical platform of the present invention. The present invention can be applied to: 1. Copyright protection, to prove the ownership and property of the video, and achieve the purpose of copyright protection without damaging the original video; 2. To determine the authenticity and integrity of the video, the watermark of the present invention is in parity Different frames, once the video is drawn (odd frames), it can be easily identified by watermark; 3. Advertisement, the embedded content can be the author's name, company logo, text with special meaning, etc., which is different from the existing form For video advertisements, the present invention embeds advertisements in the video itself, without additional equipment and expenses, and the outside world cannot eliminate the advertisement content without changing the video.

Description of drawings

Figure 1 is a flow chart of the overall scheme of visible video watermarking when the video is still.

Figure 2 is the first frame image after embedding the watermark.

Figure 3 is the second frame image after embedding the watermark.

Fig. 4 is the first frame image compressed by the H.264/AVC compression coding standard.

FIG. 5 is the second frame image compressed by the H.264/AVC compression coding standard.

(a) and (b) in Figure 6 are the U component images of the first frame of the watermarked video and the H.264/AVC compressed video, respectively.

(a) and (b) in Figure 7 are the U component images of the second frame of the watermarked video and the H.264/AVC compressed video, respectively.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

The example selects the second scheme, that is, embeds the watermark in the chrominance U component of the video. The specific implementation of the watermark is as follows:

Read the binary watermarked image and determine the image height and width.

Select the second scheme from the three schemes, that is, embed the watermark in the chrominance U component of the video. Obtain the U component of each video frame one by one, and judge the parity of the video frame. Determine the area to be embedded according to the size of the watermark image.

Determine the parity of the current video frame, and adjust the corresponding watermark value according to the parity of the video frame and the value of each pixel in the area to be embedded. The current frame is an odd frame, and the watermark is a positive (negative) value; the current frame is an even frame, and the watermark is Negative (positive) values, as in Equation 2. Embed the adjusted watermark value b into the pixels of the U component according to Equation 1. When adjusting the pixel value, it should be noted that the chromaticity range is 16-240, so when the pixel value U(i,j)≥240-b, U(i,j)=240; when the pixel value U(i,j) )≤16-b, U(i,j)=16.

Taking the next video frame as the current video frame, repeat step 3 until all video frames of the video file are processed.

In order to evaluate the effect of the present invention after embedding the watermark and the effect of the compressed watermark after compression coding, we use the method in this example to conduct experiments. The experiment uses a 352×288 foreman sequence, and the watermark information is a binary watermark image with a size of 80×46 marked with two letters of “TK”. Figure 2 and Figure 3 are the images of the first frame and the second frame after embedding the watermark respectively. Figure 4 and Figure 5 are the images of the first frame and the second frame with watermark after the video is compressed by H.264/AVC compression coding. We use the peak signal-to-noise ratio to observe the objective quality of the video after the watermark is embedded in the video, and we have made some improvements. Since the present invention uses the persistence of vision of the human eye to embed the watermark in the high frame rate video, we use the change of the pixel difference between two adjacent frames as a factor to measure the subjective quality of the video, and the watermark embedded in the peak signal-to-noise ratio The pixel values are modified as follows:

${\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; f</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; f</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; f</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; f</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$

表示经修改后带有水印的第f帧(i，j)处的像素值。在本实验中，使用352×288的foreman序列，帧数为300，因此f的可取值范围为[2,300-1]，(i，j)范围为视频帧中所有的像素值，1≤i≤352，1≤j≤288。在峰值信噪比PSNR计算过程中，我们用

替换原带有水印的视频帧。利用上述改进对foreman序列和经H.264/AVC压缩的视频序列前4帧分别进行峰值信噪比PSNR计算，结果如表1，Where X _f (i, j) represents the pixel value at the fth frame (i, j) of the original video, and X _f+1 (i, j) represents the pixel value at the (f+1) frame (i, j) of the original video Pixel value, X′ _f (i, j) represents the pixel value at the fth frame (i, j) after embedding the watermark, X′ _f+1 (i, j) represents the (f+1)th frame after embedding the watermark ( i, the pixel value at j),

Indicates the modified pixel value at frame f (i, j) with watermark. In this experiment, a 352×288 foreman sequence is used, and the number of frames is 300, so the range of possible values of f is [2,300-1], and the range of (i, j) is all pixel values in the video frame, 1≤i ≤352, 1≤j≤288. In the peak signal-to-noise ratio PSNR calculation process, we use

Replace the original video frame with watermark. Using the above improvements, the peak signal-to-noise ratio (PSNR) is calculated for the first 4 frames of the foreman sequence and the H.264/AVC compressed video sequence, and the results are shown in Table 1.

Uncompressed video PSNR Compressed video PSNR first frame PSNR=+43.74dB PSNR=+44.80dB second frame PSNR=+43.75dB PSNR=+45.09dB third frame PSNR=+43.77dB PSNR=+45.25dB The fourth frame PSNR=+43.81dB PSNR=+45.36dB

Table 1

Experimental results show that by adjusting the weight factor α, a good balance can be achieved between clearly visible watermark when the video is still and invisible when the video is playing. At the same time, after the video is compressed by H.264/AVC compression encoding, the watermark can still be clearly seen.

The present invention utilizes the characteristics of human visual persistence and the high similarity between two adjacent frames of video to embed the same bit of watermark value with opposite symbols in the same spatial space position of two adjacent frames for video with a frame rate of 60 Hz or above. It achieves the effect that the watermark is clearly visible when the video is still, but invisible when the video is playing and does not affect the subjective quality of the video. In the present invention, the watermark is still valid and visible after being compressed by the H.264/AVC compression encoding standard.

The above is only one embodiment of the present invention. As a brand-new idea and technical platform of video watermarking, the present invention is not limited to the above-mentioned embodiments. All technical solutions under the idea of the present invention belong to this invention. protection scope of the invention. It should be pointed out that for those skilled in the art, some improvements and modifications without departing from the principles of the present invention should also be regarded as the protection scope of the present invention.

From the above materials, it can be seen that the present invention has the following advantages and characteristics: the present invention proposes a brand-new idea and technical platform for video watermarking, and has the advantages of simple, effective and real-time embedding methods, and at the same time, the present invention has both visible watermarking Features and advantages of invisible and invisible watermarks: watermarks are invisible during video playback; watermarks can be intuitively visible without other operations and equipment when still; watermarks are robust and clear, resistant to noise and compression processing, and can be attacked Still clearly visible. Therefore, the present invention does not need a watermark extraction algorithm and an extraction device, and the extra cost is small. In addition, the present invention, as a brand-new technical platform, has good scalability, and technicians and developers can improve and modify it according to their own needs on the basis of the technical platform of the present invention. Based on the above advantages and characteristics, the present invention can be applied to: 1. Copyright protection, to prove the ownership and property of the video, and achieve the purpose of copyright protection without damaging the original video; 2. Determine the authenticity and integrity of the video , the watermark of the present invention is different in odd and even frames. Once the video is framed (odd frame), it can be easily identified by the watermark; 3. The embedded content of the advertisement can be the author's name, company logo, text with special meaning, etc. Different from the existing video advertisements, the present invention embeds the advertisements in the video itself, which does not require additional equipment and expenses, and the outside world cannot eliminate the advertisement content without changing the video.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (6)

1. a video watermark implementation method is characterized in that, in the opposite same position watermark value of position, the identical spatial domain of adjacent two frames insertion symbol, simultaneously to the applicable requirement of video be frame per second reach 60Hz and more than; The method is divided into odd-numbered frame and even frame with frame of video, with binary watermarking proportionally with the addition of odd-numbered frame pixel value or subtract each other, subtract each other or addition with the even frame pixel value, adjacent two frame same position pixel average are equated with the original video image pixel value, reach the effect of the image pixel value complementation of parity frame watermark embedding place.

2. video watermark implementation method as claimed in claim 1, it is characterized in that, watermark is directly embedded in the pixel of image spatial domain, watermark embedding method is: before watermark embeds, adjust watermark value b according to carrier image position to be embedded pixel value, according to the parity of frame of video, determine the positive negative value of watermark value b, such as formula 2; After watermark value is determined, the embedding grammar of watermark such as formula 1, when the watermark information position was 0, the watermark value b after carrier image pixel value and the adjustment subtracted each other; When the watermark information position is 1, the watermark value b addition after carrier image pixel value and the adjustment;

$\stackrel{\&OverBar;}{X}=\left\{\begin{array}{c}X(i,j)-b,W=0\\ X(i,j)+b,W=1\end{array}\right.$ Formula 1

Formula 2

Wherein, X (i, j) is the pixel value of video image,

Be the pixel value after the watermark embedding, W represents the watermark information position; The watermark value that parameter b indicates to embed, symbol is regulated according to the parity of the frame of video that will embed, and size is determined by pixel value; α is weight, and its value sets up on their own according to selected scheme and video; Round() function representation is exported the integer value after this function input value (α * X (i, j)) rounds up; When frame of video was odd-numbered frame, the b value was positive integer+round(); When frame of video was even frame, the b value was negative integer-round().

3. video watermark implementation method as claimed in claim 2 is characterized in that, concrete steps are:

1) reads binary bitmap;

2) the selection wish is embedded into watermark the component of frame of video, obtains one by one the frame of video of video file, judges the parity of frame of video; Determine zone to be embedded according to the watermarking images size;

3) adjust the watermark value symbol according to the frame of video parity, make the opposite in sign of the same position watermark value of adjacent two frame same positions embedding, the watermark value after adjusting is embedded in the pixel of frame of video component;

4) with next frame of video as current video frame, repeated execution of steps 3 until all frame of video of video file be disposed.

4. video watermark implementation method as claimed in claim 3 is characterized in that, described step 3) in adjust the detailed step that the watermark value symbol embeds frame of video according to the parity of frame of video and comprise: present frame is odd-numbered frame, watermark be on the occasion of; Present frame is even frame, and watermark is negative value; Watermark value after will adjusting according to formula 1 directly embeds in the pixel of frame of video component.

5. video watermark implementation method as claimed in claim 3 is characterized in that, described step 3) in adjust the detailed step that the watermark value symbol embeds frame of video according to the parity of frame of video and comprise: present frame is odd-numbered frame, and watermark is negative value; Present frame is even frame, watermark be on the occasion of; Watermark value after will adjusting according to formula 1 directly embeds in the pixel of frame of video component.

6. described video watermark implementation method as claimed in claim 3 is characterized in that described step 2) in, select wish watermark to be embedded into the component of frame of video, the selection of described component can be one or more combination of following scheme:

A. watermark is embedded in the Y component of frame of video, weight α is less than 1/8;

B. watermark is embedded in the U component of frame of video, weight α is less than 1/6;

C. watermark is embedded in the V component of frame of video, weight α is less than 1/6.

Images