JP2624507B2 - Motion compensated telecine device - Google Patents

Description

The present invention relates to an apparatus for converting a continuous image between a television image and a movie film image having different numbers of frames per unit time (hereinafter referred to as a telecine apparatus). In particular, the present invention relates to a telecine apparatus having a motion compensation type frame number conversion function by a position interpolation method.

(Summary of the Invention) The present invention relates to a telecine apparatus, in which a motion vector is detected from a movie film image in a telecine apparatus which has conventionally caused jerkiness or blurring, and a motion correction type frame number conversion using the motion vector is performed. By providing the section, the number of frames is converted by the position interpolation method so that jerkiness and blurring do not occur in the moving image portion.

(Prior Art) Generally, an 8 mm movie has a film image of 16 frames per second, a 16 mm movie or a 35 mm movie has a film image of 24 frames per second, and a television generally has an image of 30 frames (frames) per second. It is. As described above, in a conventional telecine apparatus that creates a television image from a movie image having a different number of frames per second from the television, a method of repeating the same frame several times (for example, a 2: 3 method) or Alternatively, the conversion is performed using a method of linear interpolation from several frames.

(Problems to be Solved by the Invention) The conventional telecine apparatus as described above has a drawback in that jerkiness (unnaturalness due to unsmooth motion) and blurring occur in a moving image. SUMMARY OF THE INVENTION An object of the present invention is to provide a motion-compensated telecine apparatus for performing frame number conversion of a moving image without jerkiness or blurring.

(Means for Solving the Problems) The motion compensation type telecine device according to the present invention can improve the image quality by performing interpolation so that the contour can be faithfully reproduced because the human eye mainly focuses on the contour portion. A motion vector detecting unit that detects a motion vector of an image from a movie film image, a linear interpolation interpolation type frame number converting unit, and one or more motions detected by the motion vector detecting unit. A motion correction interpolation type frame number conversion unit for performing frame number conversion by position interpolation using a vector, and a set of linear interpolation interpolation units stored in a frame memory of the linear interpolation interpolation type frame number conversion unit; A signal stored in one or a plurality of frame memories of the motion correction type frame number conversion unit and a signal obtained by performing motion correction using one or a plurality of sets of motion vectors; And an output image selection unit for selecting and switching output each time.

(Operation) Since the image of the frame to be interpolated is created using the position interpolation method using motion vectors, the outline of the moving image part can be faithfully reproduced, eliminating jerkiness and blurring and significantly improving the image quality of the moving image. it can.

(Embodiment) Hereinafter, an embodiment of a motion compensation type telecine device according to the present invention will be described with reference to the drawings.

FIG. 2 (a) shows a signal F1 of a _first frame and a signal F1 of a second frame of a movie film image (hereinafter abbreviated as movie image).
_2, a diagram showing a time relationship between the signal F _a piece of television image to interpolate (hereinafter abbreviated as TV image) between those pieces, shows taking time on the horizontal axis. Movie image 1
There are from the time of the frame pieces to insert out of the TV image at the time that has passed through the time t _1, and from the time of the frame of the television image to be interpolated to the time of the second frame of the movie image is assumed to go through the time t _2, the inner The time of the frame of the TV image to be inserted is the first of the movie image
The time position of the piece and the second piece It is in the position divided internally.

For example, when creating a television image of 30 frames per second from a 16 mm movie image of 24 frames per second, if one of the two images of the movie image and the television image is matched, the above-described internal division ratio is used. Is 4/5, 3/5, 2/5, 1/5 or 0.

FIGS. 2 (b) and 2 (c) are block diagrams showing the operation of the frame number conversion method based on the linear interpolation interpolation method and the position inner layer method using a motion vector, respectively. . The multiplier constants k ₁ and k ₂ used in both figures are When t ₁ = t ₂ , k ₁ = k ₂ = 1/2. Second
If you replace the one to zero in the linear interpolation in the interpolation formula (1) and (2) of t ₁ or t ₂ shown in FIG. (B),
The same frame is repeatedly interpolated several times.

In either frame number conversion method described above, the signal F _a for each pixel in the frame (frame) of the television image to interpolated, calculated from the signal F ₁ and F ₂ of each pixel of the first and second frames of the movie image and it is designed to calculate the signal F _a for all the pixels of the frame of the television picture.

In the linear interpolation method shown in FIG. 2B, the first and second movie images at the same position as the currently calculated pixel of the frame of the television image to be interpolated (hereinafter referred to as the frame to be interpolated) are used. Input the signals F ₁ and F ₂ to the frame pixel, and interpolate the frame pixel signal
Output F _a and the frame difference. That is, the signal F ₁ of the pixels in the first frame of the same position as the pixel of interest signal calculation frame interpolating is sent to the multiplier 11 and subtractor 20, signals of the pixels of the second frame of the same position F ₂ Is sent to the multiplier 12 and the subtractor 20. Multiplier 11 is a constant k ₁ from constant recorder 41 (according to equation (1))
Is multiplied by the signal F ₁ of the first frame and sent to the adder 30,
Is the constant k ₂ (according to equation (2)) from the constant storage 42 and the second
Is multiplied by the signal F ₂ of the frame and sent to the adder 30. The adder 30 outputs a signal F _a frame interpolating adds the output of multiplier 11 and 12, the subtracter 20 is the absolute of the difference between the signal F ₂ of the signal F ₁ and the second frame of the first frame The value is calculated and output as a frame difference.

In the position interpolation method using the motion vector shown in FIG. 2C, a motion vector is separately detected from the first frame and the second frame of the movie image, and the first frame and the second frame are separately detected. Piece signal F ₁
And F ₂ , and t ₁ / (t ₁ + t ₂ ) for the motion vector
Against a _{-t 2 (ti 1 + t 2} ) type are multiplied by, but outputs the signal F _a frame interpolating the motion compensation frame difference, at this time, the position of the pixel of the frame to be interpolated, First input
And the position of the pixel of the second frame is ₁ t ₁ / (t ₁ + t ₂ ) and −
is controlled by _{t 2 / (t 1 + t} 2).

That is, the motion correction block 51 receives t ₁ / (t ₁ + t ₂ ) and calculates −t ₁ / (t ₁₎ from the position of the pixel to be interpolated for the frame to be interpolated.
+ T ₂ ), selects the signal F ₁ of the pixel of the first frame at the position shifted by + t ₂ , and sends it to the multiplier 11 and the subtractor 20 to obtain the motion compensation block.
52 receives −t ₂ / (t ₁ + t ₂ ) and converts the signal F ₂ of the pixel of the second frame at a position shifted by t ₂ / (t ₁ + t ₂ ) from the position of the pixel symmetrical to the operation of the frame to be interpolated. Select and multiplier 12 and subtractor
Send to 20.

The multiplier 11 multiplies the constant k ₁ (according to equation (1)) from the constant storage 41 by the output of the motion compensation block 51 and sends the result to the adder 30. The multiplier 12 outputs the constant k ₂ from the constant storage 42. (Equation (2)
) And the output of the motion compensation block 52.
Send to 30. The adder 30 outputs a signal F _a pixel of the frame to be interpolated by adding the outputs of multipliers 11 and 12 (frame), the subtracter 20 calculates the absolute value of the difference between the motion compensation block 51 and 52 And outputs it as a motion compensation frame difference.

The pixel of the first frame or t ₂ / (t ₁ + t ₂ ) is located at a position shifted by −t ₁ / (t ₁ + t ₂ ) from the pixel of the interpolation frame whose operation is symmetric. The second in the shifted position
If there is no pixel of the frame No., a different processing method is used, but this is not the subject of the present invention, and a detailed description thereof will be omitted.

In general, for a pixel relating to an object in which only the relevant motion vector has moved during the time interval between the time when the first frame and the second frame have been photographed, that is, t ₁ + t ₂ hours, the pixel moves as long as the signal value does not change within that time. The outputs of the correction blocks 51 and 52 are the same, and therefore the output signal _Fa is also the same. t moving objects and in accordance with different motion vector with the motion vector to ₁ + t ₂ hours, the pixels of the object did not move to t ₁ + t ₂ hours, the output of the motion compensation blocks 51 and 52, so that the output signal
_Fa may coincide or differ.

FIG. 1 is a block diagram of one embodiment of a motion compensation type telecine apparatus according to the present invention. First, a movie film image is scanned by a film image reading unit 1, and the read signal is sent to an A / D converter 2 and converted into a digital signal.

The motion vector detection section 4 detects a vector by comparing the signal of each pixel of the current frame and the previous frame output from the frame memory 3. The motion vector indicates the direction and distance that the object has moved between one frame of the film when the image includes a moving object. As a method of detecting a motion vector, there are known several methods such as a pattern matching method using a representative point and a gradient method, in addition to a block matching method disclosed in Japanese Patent Application Laid-Open No. 61-269475. Either of these methods may be used in the motion vector detection unit 4 in the present invention.

One or a plurality of motion vectors _n (up to about 4) detected by the motion vector detection unit 4 are sent to the frame order control unit 5, and are fixed according to the temporal position of the frame (frame) to be inserted. After that, it is sent to the motion-compensation interpolation type frame number conversion unit 7 in which a plurality of frame memories (71 to 7n) for motion-compensation interpolation type signals are incorporated for each motion vector _n .

As described with reference to FIG. 2 (b), the linear interpolation interpolation type frame number conversion section 6 incorporating the frame memory 61 outputs the signal of the pixel of the first and second frames at the same position as the pixel position of the frame to be inserted.
Sequentially enter F1 and F _2, signals of all pixels of the frame to be interpolated F _a
And the frame difference is calculated.

The motion-compensated interpolation type frame number conversion unit 7, as described in FIG. 2 (c), the signal F ₁ and F ₂ of the pixels of the first and second piece at a position corresponding to each motion vector _n sequentially Enter
Signal F _a and the motion compensation frame difference of all pixels of the frame to be interpolated for each motion vector _n is calculated.

The output image selection unit 8 includes a linear interpolation interpolation type frame number conversion unit 6
The signal obtained by performing the motion compensation by the number _n of the motion vectors and the motion-compensated signal outputted from the motion-compensation-interpolation-type frame number conversion unit 7 output from It selects and switches for each pixel using the minimum method and outputs.

FIG. 3 is an explanatory diagram of the frame difference addition value minimizing method, in which the vertical axis represents time, and the horizontal axis represents pixel positions. Although the positions of the pixels are shown in one dimension, there is no problem assuming that they represent two dimensions. It assumes the position C of the calculation target pixel signal F _a frame interpolating a reference. If the pixel is actually a pixel of an object that moves according to the motion vector ₁ , then at the first frame t ₁ hours ago, At position A, and this signal Expressed by t In the _second piece _two hours later, At position B, and this signal Expressed by As a signal F _a point C But as the frame difference Is output.

For example, when the calculation is performed in accordance with the fake motion vector ₂ , the position C is determined as follows in the first frame t ₁ hours before. At the position D of the second piece after t ₂ hours At position E of Expressed by As a signal F _a point C But as the frame difference Is output.

Signals at positions A, B, C, D, and E are FA, FB, FC,
If represented by FD and FE, the motion vector at position C
Frame difference for a frame difference L ₁ and ₂ about ₁
L ₂ = respectively _{L 1 | FA-FB | L} 2 = | FD-FE | become. In this case, there is no problem if L ₁ = 0 and L ₂問題 0, but there are cases where L ₁ = L ₂ = 0 depending on the picture. In that case, the motion vector at the position C cannot be specified.

The case where the minimum frame difference addition value method is used will be described. Considering the sum of the motion vector ₂ For example, the first position D signal frame by the motion vector ₂ with respect to the reference position C And a signal at the position G of the second frame moved from the position D according to the motion vector _1. Let the absolute value of the difference be the sum. Considering the added value for the p-th motion vector _p instead of motion vector ₂ , the position of the first piece Signal and the position of the second piece Is the added value.

This can be changed to the description of the frame difference addition value L _r 'for the r-th motion vector _r. Therefore, as shown in FIG.
Applying to the case where only ₁ and ₂ exist,
Frame difference addition value L ₁ for ₁ ', the sum of the absolute value of the difference between the position of the signal coupled by the solid line in Figure 3, , And the frame difference addition value L ₂ for the ₂ ', 3
In the figure, the sum of the absolute value of the difference between the signals At the C position of the frame to be interpolated, a signal calculated based on the motion vector ₁ is selected.

In this manner, all the pixels of the frame to be inserted in the output image selection unit 8, picked those minimum frame difference sum value, and outputs as a signal F _a of each pixel of the output image.

(Effect of the Invention) The conventional telecine apparatus has a drawback that jerkiness and blur occur in a moving image portion. However, the motion-corrected telecine apparatus according to the present invention has a motion-correction type using a motion vector detection function and a position interpolation method. By providing the frame number conversion unit, a converted image without jerkiness or motion blur can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a telecine apparatus according to the present invention. FIG. 2 (a) is a diagram showing a time relationship between a signal of a first frame, a signal of a second frame, and a signal of a frame to be interpolated. 2 (b) and (c) are block diagrams showing the operation of the frame number conversion method by the linear interpolation interpolation method and the motion compensation position interpolation method, respectively. FIG. 3 is an explanatory diagram of the frame difference addition value minimization method. . 1 Film image readout unit 2 A / D converter 3 Frame memory 4 Motion vector detection unit 5 Frame order control unit 6 Linear interpolation interpolation type frame number conversion unit 7 Motion Correction interpolation type frame number conversion unit 8 Output image selection unit 11,12 Multiplier 20 Subtractor 30, Adder 41,42 Constant storage unit 51,52 Motion correction block 61,71,72 …… Frame memory

Claims (1)

(57) [Claims] 1. A telecine apparatus for converting a film image of a movie into a television signal, a motion vector detecting unit for detecting a motion vector of the image from the film image,
A linear interpolation interpolation type frame number conversion unit, a motion correction interpolation type frame number conversion unit that performs frame number conversion by a position interpolation method using one or a plurality of motion vectors detected by the motion vector detection unit, A set of linear interpolation interpolation signals stored in a frame memory of the linear interpolation frame number conversion unit and a signal stored in one or a plurality of frame memories of the motion compensation type frame number conversion unit are An output image selection unit for selecting and switching and outputting each pixel from a signal motion-corrected by one or a plurality of sets of motion vectors using a frame difference addition value minimization method. Type telecine device.