JP2533538B2 - Image signal frame number conversion method - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image signal frame number conversion system for converting the number of frames among system conversions for converting a scanning system of a television image signal, and more particularly It is intended to obtain a converted output image signal with less deterioration of image quality such as judder in which definition deterioration due to movement and vertical lines of an image become jagged.

(Prior Art) As a conventional image signal frame number conversion method of this kind,
A linear interpolation method that uses an interpolation filter to compensate for the shortage of image frames caused by the difference in the number of frames per second of the image signal, or the position shift of the interpolated frame image due to the motion by detecting the motion vector that represents the motion of the image content. Although the position interpolation method for correcting is used, each of the conventional methods has the following problems.

(Problems to be Solved by the Invention) That is, in the linear interpolation by the linear interpolation method,
The image signal of the interpolated frame is an image signal of a signal level obtained by weighted averaging the image signal levels of the front and rear frames by an interpolation ratio that is inversely proportional to the distance from the frame before and after the interpolated frame. Signal every second
In the case of conversion into an image signal of 50 frames, as shown in FIG. 2, when the timing of the first frame is matched between the converted input / output image signals as a timing reference, for example, the conversion output image signal For 3 frames, the converted output signal level is obtained as the sum of 60% of the signal level of the 3rd frame and 40% of the signal level of the 4th frame in the converted input image signal.

Therefore, if the signal waveforms of the conversion input third frame and the fourth frame have a ladder shape as shown in FIGS. 3A and 3C, respectively, the signal waveform of the conversion output interpolation frame becomes As shown in (b) of the figure, the rising and falling are deteriorated by the weighted average of the rising and falling in the preceding and following frames, and when the image movement is irregular, so-called motion blur with reduced definition occurs, When the image movement is uniform, the weighted averaging is performed irregularly to cause judder.

On the other hand, in the position interpolation method, the interpolation frame is calculated based on the motion vector of the image detected by pattern matching using the direction and amount of the positional deviation of the image parts of the same pattern between two adjacent frames of the converted input image signal. After correcting the positions of the frames before and after, the sum of the image signals of the frames before and after is corrected by the coefficient value 1/2 to obtain the converted output image signal of the interpolated frame. Therefore, as shown in FIG. 4, when the motion vector V is detected from the positional deviation in the x direction between the preceding frame and the succeeding frame which are adjacent to each other in the converted input image signal, the motion of the image varies or If there is a lot of noise, the detection accuracy of the motion vector decreases, and as a result of performing the position correction based on the erroneously detected motion vector, worm-like disturbance occurs in the converted output image, and the image quality deteriorates.

As described above, the conventional image signal frame number conversion method has a problem that some image quality deterioration occurs.

(Means for Solving Problems) An object of the present invention is to solve the above-mentioned conventional problems, and to make a motion blur due to an influence of a motion state of an image signal, noise, or the like.
An object of the present invention is to provide an image signal frame number conversion method that does not cause image quality deterioration such as judder or worm-eaten.

That is, in the image signal frame number conversion method of the present invention, a Fourier transform is performed for each frame on a two-dimensional image signal composed of spatial rectangular coordinate variables, and each term of the converted output function is converted into a polar coordinate function composed of an amplitude term and a phase term. Then, by subjecting each amplitude term and each phase term of the polar coordinate function to interpolation respectively, and performing the inverse Fourier transform on each amplitude term and each phase term of the interpolated output, and transforming into a spatial Cartesian coordinate variable, A two-dimensional image signal having a different number of frames per second from the two-dimensional image signal is formed.

(Operation) Therefore, according to the present invention, there is no deterioration in image quality such as motion blur, judder, and worm bite in the output image converted by the number of frames, and further, the labor for detecting the motion vector of the image in advance in the conversion of the number of frames is omitted. be able to.

(Example) Hereinafter, the present invention will be described in detail with reference to the drawings.

First, FIG. 1 shows a basic configuration of an image signal frame number conversion apparatus according to the present invention in the case of converting an image signal having a frame number N per second into an image signal having a frame number M per second.

The illustrated frame number conversion device applies a Fourier transform to a two-dimensional image signal consisting of spatial rectangular coordinate variables for each frame, and also transforms each term of the transform output function into a polar coordinate function consisting of an amplitude term and a phase term, respectively. Part 1, interpolating parts 2 and 3 for performing linear interpolation and the like on each amplitude term and each phase term of the transformed output, and performing inverse Fourier transform on each amplitude term and each phase term of the interpolating output It is basically configured by a Fourier inverse transform unit 4 which forms a frame number conversion output two-dimensional image signal composed of spatial Cartesian coordinate variables.

A converted output image obtained by supplying the Fourier transform unit 1 with a converted input image signal of the number of frames per second N in which the horizontal position in the image frame is x, the vertical position is y, and the frame order is n. The signal F _n (X, Y) is given by the following equation (1).

Where X: horizontal spatial frequency Y: vertical spatial frequency []: Fourier transform symbol r _n (X, Y): amplitude term θ _n (X, Y): phase term Fourier transform and coordinate transform are performed as described above. Amplitude term r _n (X, Y) and phase term θ of the generated image signal F _n (X, Y)
_By supplying _n (X, Y) to the interpolation units 2 and 3, respectively, and appropriately performing interpolation such as linear interpolation, Lagrange interpolation, and spline function interpolation by an interpolation filter, the number of frames per second N Convert the number of frames between M and
A transformed output image signal consisting of spatial polar coordinate variables consisting of the amplitude term r _i (X, Y) and the phase term θ _i (X, Y) is obtained. Then, since the interpolation is performed in the form of the spatial polar coordinate variable, it means that the interpolation is performed including the positional deviation due to the movement of the image between the adjacent frames. Then, the interpolation interpolation output amplitude term r _i (X, Y) and phase term θ _i (X, Y) of the interpolating sections 2 and 3 are both supplied to the inverse Fourier transform section 4 to perform the inverse Fourier transform. Then, coordinate conversion between polar coordinates and rectangular coordinates is performed according to the following equation (2).

Here, ^-1 []: Fourier inverse transform symbol Next, FIG. 5 shows an example of a specific configuration of the frame number conversion apparatus of the present invention which basically operates as described above. In the configuration shown in the figure, linear interpolation is performed using interpolation filters 5 and 6 and decimation circuits 7 and 8 as the interpolation units 2 and 3, and the image signals of the number N of frames per second are generated by the interpolation filters 5 and 6. It is converted into an image signal having the number of frames per second which is the least common multiple of the number of input / output frames N and M, and is further converted into an image signal of the number of frames M per second by the thinning circuits 7 and 8. In the example shown in FIG. Therefore, if the interpolation outputs of the interpolation filters 5 and 6 obtained every 1/300 seconds are taken out through the thinning circuits 7 and 8 every 1/50 seconds, a required image signal of 50 frames per second can be obtained.

Next, as another example of the specific configuration of the frame number conversion apparatus according to the method of the present invention, a configuration in which the circuits 9 and 10 for performing the simplest linear interpolation among linear interpolations are used in the interpolation units 2 and 3. FIG. 6 shows the aspect of the frame number conversion by the configuration shown in FIG. 7 when the image represented by the two-dimensional image signal is translated in the x direction as shown in FIG. An example in which the signal level of an image signal regarded as a signal is changing will be described.

In this case, the following expression (4) is established between the two-dimensional one-dimensional image signals of two adjacent frames.

f ₁ (x) = α · f ₀ (X−Δx) (4) where f ₁ (x): current frame image signal f ₀ (X−Δx): previous frame image signal α: rate of change in signal level Δx: Amount of movement in x direction When the transformed input image signal is subjected to Fourier transform to transform the transformed output function into the polar coordinate amplitude term and phase term, the following equation (5) is established.

When the above-mentioned amplitude term and phase term are linearly interpolated, the following equation (6) is established.

Here, k ₁ and k ₂ are interpolation ratios, respectively, and k ₁ + k ₂ = 1 (7) in the relationship of the following equation (7). Equation 8 is established.

Therefore, as shown by the dotted line in the middle of the previous frame image signal shown by the solid line and the current frame image signal shown by the one-dot chain line in FIG. 7, the position interpolated frame image signal including the positional shift due to the motion of the image is the frame. It is obtained as a number conversion output.

However, in the position interpolation that performs motion correction using the conventional motion vector, it becomes difficult to accurately detect the motion vector for the image signal whose signal level changes with the motion of the image as described above. As a result, accurate position correction could not be performed, but the problem disappears according to the method of the present invention.

(Effect of the Invention) As is apparent from the above description, in image signal frame number conversion, since linear interpolation or motion-correction type position interpolation has been conventionally performed, image blurring such as motion blur and judder occurs. Alternatively, although the image quality of the converted output signal is deteriorated due to the erroneous detection of the motion vector, according to the present invention, the position interpolation including the position shift of the image due to the motion of the image is performed without detecting the motion vector in advance. Can be performed, and therefore motion blur, judder, or
It is possible to achieve conversion of the number of image signal frames without causing worm-eating-like image quality deterioration due to motion vector erroneous detection.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic configuration of an image signal frame number converter according to the present invention, FIG. 2 is a diagram showing a mode of image signal frame number conversion, and FIGS. 3 (a) to 3 (c). FIG. 4 is a signal waveform diagram sequentially showing the mode of image quality deterioration associated with the conventional linear interpolation. FIG. 4 is a signal waveform diagram showing the mode of the conventional motion compensation type position interpolation. FIG. 5 is an image according to the method of the present invention. FIG. 6 is a block diagram showing an example of a specific configuration of the signal frame number conversion device, FIG. 6 is a block diagram showing another example of the same configuration, and FIG. 7 is an image signal frame number conversion by the method of the present invention. It is a signal waveform diagram which shows the example of an aspect. 1 ... Fourier transform unit 2,3 ... Interpolation unit 4 ... Fourier inverse transform unit 5,6 ... Interpolation filter 7,8 ... Thinning circuit 9,10 ... Linear interpolation circuit

Claims (1)

(57) [Claims] 1. A two-dimensional image signal consisting of spatial Cartesian coordinate variables is subjected to a Fourier transform for each frame, each term of the transform output function is transformed into a polar coordinate function consisting of an amplitude term and a phase term, and each of the polar coordinate functions is transformed. By performing interpolation interpolation on the amplitude term and each phase term, and performing inverse Fourier transform on each amplitude term and each phase term of the interpolation interpolation output, and transforming into a spatial rectangular coordinate variable, the two-dimensional image signal is An image signal frame number conversion system characterized in that a two-dimensional image signal having different number of frames per second is formed.