CN102014267B - Subtitle area detection method - Google Patents

Abstract

A method for detecting a caption area. First, caption pixel detection is performed on each target pixel of the current scan line to determine whether each pixel of the current scan line is a caption pixel. Counting the number (spf) of caption pixels of the previous frame, and setting the caption pixels of the current scanning line as static pixels when the number (spf) of the caption pixels of the previous frame is greater than a preset value. Counting the number (spl) of caption pixels of a previous scanning line of the current frame, and adjusting and setting partial non-static pixels of the current scanning line as static pixels according to the previous frame and the next frame when the number (spl) of caption pixels of the previous scanning line of the current frame is greater than a preset value.

Description

Subtitle area detection method

technical field

The present invention relates to de-interlacing conversion, in particular to a subtitle region detection method for de-interlacing.

Background technique

The general TV broadcast video signal adopts interlaced format, such as NTSC, PAL, SECAM, which are sequentially displayed on the TV screen with odd field (field) and even field, and uses the principle of visual persistence to display the frame ( frame) of the video content. Interlaced video signals can deliver acceptable video quality using only low bandwidth; however, the disadvantages are reduced vertical resolution and flickering of lines or areas. The video signal of a general computer monitor adopts a non-interlaced or progressive format, which directly displays the video content of the frame on the computer monitor.

In order for an interlaced video signal to be displayed on a display with a sequential signal format (such as a computer monitor), the interlaced video signal must first be converted into a non-interlaced/sequential video signal. This format conversion process is called de-interlacing. ) or octave sequentially. By means of de-interlacing conversion, the original odd field and even field are combined to generate a frame.

The methods for de-interlacing conversion of video signals can be classified into the following two types: spatial conversion and temporal conversion. In spatial transformation, only pixels in the same field are used to generate new pixels, so this transformation is generally also called intra-field transformation. In temporal conversion, adjacent pixels of different fields are used to generate new pixels, so this conversion is also generally called inter-field conversion. Usually, the static area is de-interleaved by temporal or inter-field interpolation techniques, while the moving area is de-interleaved by spatial or intra-field interpolation techniques.

Generally speaking, motion detection (motion detection) is usually judged by the difference between corresponding pixels of the same parity field (sameparity field); a larger difference indicates that there is movement, and a smaller difference indicates that there is no movement. However, if the background of the subtitle area moves, the subtitle area may be regarded as a moving area for de-interlacing, thus causing flickering of the subtitle area.

In view of the fact that the traditional technology cannot correctly determine the subtitle area and thus cause the subtitle area to flicker, it is urgent to propose a novel motion detection and de-interlacing mechanism to correctly detect the subtitle area so as to facilitate de-interlacing.

Contents of the invention

In view of the above, an embodiment of the present invention proposes a subtitle region detection method for correctly detecting the subtitle region for de-interlacing and improving the flickering phenomenon of the subtitle region.

According to an embodiment of the present invention, firstly, subtitle pixel detection is performed on each target pixel of the current scanning line to determine whether each pixel of the current scanning line is a subtitle pixel. The number of subtitle pixels in the previous frame (spf) is counted, and when it is greater than a preset value, the subtitle pixels of the current scanning line are set as static pixels. Count the number of subtitle pixels (spl) of the previous scan line of the current frame, and when it is greater than a preset value, adjust and set some non-stationary pixels of the current scan line according to the previous frame and the next frame Still pixels.

According to an embodiment, the above subtitle pixel detection step includes the following steps. First, the motion detection of the target pixel is performed. Detect the brightness values of adjacent pixels of the target pixel, and determine whether there is a particularly large or small brightness value. Next, it is checked whether the state of the adjacent pixel corresponding to the target pixel is static in the previous frame. Finally, detect whether the target pixel is located at the boundary of the subtitle.

Description of drawings

Figure 1 illustrates partial scan lines of three consecutive frames.

FIG. 2 shows a flowchart of a subtitle pixel detection method.

FIG. 3 shows a flow chart of a subtitle area detection method according to an embodiment of the present invention.

Figure 4 illustrates a scanline with subtitles and its corresponding previous and next frame scanlines.

[Description of main component symbols]

21-24 Subtitle pixel detection steps

31-38 Subtitle area detection steps

PA, PC, PE The scan line of the previous frame

B, D Scanning line of the current frame

NA, NC, NE The scan line of the next frame

pa, pb, pc, pd, pe Pixels of the previous frame

b, d Pixels of the current frame

na, nc, ne Pixels of the next frame

Detailed ways

FIG. 1 illustrates partial scan lines of three consecutive frames (ie, previous frame, current frame and next frame). Wherein, the previous frame and the next frame are odd fields, and the current frame is an even field. The previous frame displays scan lines PA, PC, PE, the current frame displays scan lines B, D, and the next frame displays scan lines NA, NC, NE.

FIG. 2 shows a flow chart of a subtitle pixel detection method, which is used to detect a current pixel t to be de-interlaced (also referred to as a target pixel) to determine whether it may be a subtitle pixel. Although the present embodiment uses steps 21 to 24 to sequentially detect the target pixel t, however, the order of these steps can be changed. Furthermore, some steps may be omitted, and other additional steps may be added.

Firstly, the motion detection (motion detection) of the target pixel t is performed in step 21 . Given that subtitles generally last for several seconds, it is convenient for viewers to read them. Therefore, when performing motion detection on the subtitle area, the difference between at least two fields with the same parity will be extremely small. In this embodiment, the above-mentioned difference between the odd and even fields is obtained by measuring multiple pixels of the previous frame corresponding to the target pixel t (for example, multiple pixels in a window covering the pixel pc) and multiple pixels of the next frame. The sum of absolute differences (SAD) between pixels (for example, multiple pixels covering a window of pixel nc). The value SAD of the sum of absolute differences can be expressed by the following formula:

$\mathrm{<span\; class="notranslate">\; SAD</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; nc</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; pc</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span>$

Among them, nc _i represents the pixels on the scan line NC of the next frame, pc _i represents the pixels on the scan line PC of the previous frame, and k represents the number of pixels in the window.

If the sum SAD of the above-mentioned absolute differences is less than a preset value T, then indicate that the state of the target pixel t is static (static), and continue the subsequent steps; if the sum SAD of the absolute differences is greater than the preset value T, then indicate as move (motion), and end the process of FIG. 2 and determine that the target pixel t is not a subtitle pixel.

Next, detect the adjacent pixels of the target pixel t in step 22 (for example, the pixel b above and below the target pixel t in the current frame, and the pixel pc corresponding to the target pixel t in the previous frame) The brightness value is used to detect whether there is a particularly large or small brightness value. In order to make the subtitle stand out and distinguish it from the background, the subtitle usually uses a dark (especially small brightness value) outer frame and a bright color (especially large brightness value) inner text to enhance the contrast. Therefore, when a particularly large or small luminance value is detected in step 22, the target pixel t may be located in the subtitle area, and the subsequent steps are continued; otherwise, the process of FIG. 2 is ended and the target pixel t is determined not to be a subtitle pixel. In this embodiment, step 22 thus determines whether the pixels b, d, pc have particularly large brightness values (greater than the preset value T2) or have particularly small brightness values (less than the preset value T3). That is, b>T2 or d>T2 or pc>T2 or b<T3 or d<T3 or pc<T3.

However, if the pixel falls in the smooth region (that is, the general monotone luminance value), the above steps 21 and 22 are not enough to distinguish the subtitle. Even, there may be a lot of subtitle pixels detected by mistake. Therefore, this embodiment continues with steps 23 and 24 for further detection.

In step 23, it is checked whether the state of the adjacent pixels corresponding to the target pixel t (eg, pixels pb and pd) in the previous frame is still. If yes, proceed to subsequent steps; otherwise, end the process of FIG. 2 and determine that the target pixel t is not a subtitle pixel. In this embodiment, it is checked whether the states of the pixels pb and pd are both stationary.

Finally, in step 24, it is detected whether the target pixel t is located at the boundary of the subtitle. The boundary of the subtitle generally has one of the following two situations: (a) the (absolute) difference between the upper and lower adjacent scan lines (such as scan line B and scan line D) of the target pixel t is very large; (b) the target pixel t The (absolute) difference between the upper and lower adjacent scan lines of t is very small, but the (absolute) difference between the upper and lower adjacent scan lines and the scan line (such as scan line PC) corresponding to the target pixel t in the previous frame The difference is huge. If there is case (a) or (b), then determine that the target pixel is a subtitle pixel; otherwise, end the process of FIG. 2 and determine that the target pixel t is not a subtitle pixel. In this embodiment, the case (a) is to determine whether the absolute difference between the window average value of scan line B (covering pixel b) and the window average value of scan line D (covering pixel d) is greater than a preset value T4. That is, |B-D|>T4. Case (b) is to determine whether the absolute difference between the window mean value of scan line B (covering pixel b) and the window mean value of scan line D (covering pixel d) is less than a preset value T5 (that is, |B-D|< T5), and whether the absolute difference between scan line B/scan line D and scan line PC (covering pixel pc) is greater than the preset value T4 (that is, |B-PC|>T4 or |D-PC|>T4) .

FIG. 3 shows a flow chart of a subtitle area detection method according to an embodiment of the present invention, and subsequent (not shown in the flow chart) de-interlacing can be performed according to the obtained motion detection results. Firstly, in step 31 , subtitle pixel detection is sequentially performed on each target pixel of the current scanning line according to the process shown in FIG. 2 , so as to determine whether each pixel of the scanning line is a subtitle pixel.

Then, in step 32, the number of subtitle pixels in previous frame (subtitle pixels in previous frame, spf) is counted. Since the same subtitle will appear continuously in multiple frames, it can be known whether the current frame has subtitles according to the result of the previous frame. Generally speaking, when the number of subtitle pixels in a frame (spf) is large enough, it is possible to actually have subtitles.

When the number (spf) of the subtitle pixels of the previous frame is greater than a preset value T5 (step 33), then the subtitle pixels of the current scanning line are set as static (static) pixels (step 34); otherwise, promptly according to general The results of the motion detection (step 35) are used for de-interlacing. For example, temporal or inter-field interpolation techniques are used to perform de-interleaving of stationary areas, while spatial or intra-field interpolation techniques are used to perform de-interleaving of moving areas. Then, in step 36, the number of subtitle pixels in previous scan line (subtitle pixels in previous scan line, spl) is counted (current frame).

If the number of subtitle pixels (spl) of the previous scanning line is greater than a preset value T6 (step 37), then the current scanning line is a scanning line with subtitles; otherwise, according to the general motion detection (step 35) The results are deinterleaved. According to this embodiment, whether the current frame has subtitles can be known from the calculated data (such as spf) of the previous frame. In order to save storage space for additional recording information, the position of the subtitle area is obtained from the data of the current frame (such as spl).

Among the scanlines determined to have subtitles, some pixels are still and some pixels are not. If the absolute difference between the pixels of the previous frame and the pixels of the next frame corresponding to the “non-stationary” pixel is smaller than a preset value T7, then it is adjusted and set as a static pixel in step 38 . Figure 4 illustrates a scanline with subtitles and its corresponding previous and next frame scanlines. In this example, the third pixel t3 of the scan line with subtitles is a non-stationary pixel, the corresponding pixel of the previous frame is pc3, and the corresponding pixel of the next frame is nc3. If the absolute difference between the pixel nc3 and the pixel pc3 is smaller than the preset value T7 (ie, |nc−pc|<T7), then the pixel t3 is adjusted to be a static pixel.

According to this embodiment, the subtitle area and its adjacent area can be determined as a still area, so that its motion detection is not affected by the background movement, and the still area can be implemented by temporal or inter-field interpolation technology to de-interleave.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the scope of the appended claims of the present invention; any other equivalent changes or modifications that do not deviate from the spirit disclosed by the invention should be included in within the scope of the appended claims.

Claims (8)

1. A subtitle area detection method, comprising: Perform subtitle pixel detection on each target pixel of the current scan line to determine whether each pixel of the current scan line is a subtitle pixel; Count the number of subtitle pixels in the previous frame (spf); When the number (spf) of the subtitle pixels of the previous frame is greater than a preset value, the subtitle pixels of the current scanning line are set as static pixels; Count the number of subtitle pixels (spl) of the previous scan line of the current frame; and When the number of subtitle pixels (spl) of the previous scan line is greater than a preset value, if the absolute difference between the pixels of the previous frame and the pixels of the next frame corresponding to the non-stationary pixels of the current scan line is less than a default value, adjust and set the non-stationary pixels of the current scan line as static pixels. 2. The subtitle area detection method according to claim 1, wherein the above-mentioned subtitle pixel detection step comprises the following steps: perform motion detection of the target pixel; and Detecting the luminance values of adjacent pixels of the target pixel to determine whether there is a particularly large or particularly small luminance value. 3. The subtitle region detection method according to claim 2, wherein the motion detection step of the above-mentioned target pixel comprises: Measuring the sum of absolute differences (SAD) between a plurality of pixels of the previous frame corresponding to the target pixel and a plurality of pixels of the next frame; Wherein, if the sum of absolute differences (SAD) is less than a preset value, it indicates that the state of the target pixel is stationary; otherwise, it indicates that it is moving. 4. The subtitle region detection method according to claim 2, wherein said adjacent pixels include pixels above and below the target pixel in the current frame, and pixels corresponding to the target pixel in the previous frame . 5. subtitle region detection method according to claim 2, wherein above-mentioned subtitle pixel detection step also comprises: Check whether the state of the adjacent pixel corresponding to the target pixel is static in the previous frame. 6. The subtitle region detection method according to claim 5, wherein the above-mentioned subtitle pixel detection step further comprises: Detect whether the target pixel is located at the boundary of the subtitle. 7. The subtitle area detection method according to claim 6, wherein the detection step of the above-mentioned subtitle boundary comprises: Determine whether the absolute difference between the average value of multiple pixel points corresponding to the target pixel on the previous scan line of the current scan line and the average value of multiple pixel points corresponding to the target pixel on the next scan line is greater than a predetermined value set value. 8. The subtitle area detection method according to claim 6, wherein the detection step of the above-mentioned subtitle boundary comprises: Determine whether the absolute difference between the average value of multiple pixel points corresponding to the target pixel on the previous scan line of the current scan line and the average value of multiple pixel points corresponding to the target pixel on the next scan line is less than a predetermined value Set a value, and the average value of multiple pixel points corresponding to the target pixel on the previous scan line or the next scan line is the average value of multiple pixel points corresponding to the target pixel on the corresponding scan line of the previous frame Whether the absolute difference of the values is greater than a preset value.

Images