TWM628629U - Stereo image generating device - Google Patents

Abstract

A stereo image generating device is provided, which performs: obtaining a first depth information map of a first image, each pixels of the first depth information map has a corresponding depth information; performing uniformity process on a plurality of edge pixels which are within a predetermined width from a plurality of edges of the first depth information map, to make the processed edge pixels have the same depth information, to establish a second depth information map; setting a pixel shifting amount corresponding to each pixels of the first image based on the depth information corresponding to each pixels of the second depth information map; performing pixel shifting operation on the first image to generate a second image; and displaying the first image and the second image of a stereo image.

Description

Stereoscopic image generation device

The present invention relates to an image processing device and an image processing method, in particular to a stereoscopic image generating device.

Among the conventional stereoscopic image generation methods, there is a method using a monocular depth estimation technique.

FIG. 1 illustrates the above-mentioned conventional stereoscopic image generation method. First, the method uses a convolutional neural network model to estimate the depth information of each pixel of the first image 21 (that is, the original image used as the first viewing angle) to obtain the first depth information map 22 . Then, the method sets the pixel offset 23 of each pixel in the first image 21 according to the depth information corresponding to each pixel in the first depth information map 22 . Finally, the method uses the pixel offset 23 to perform pixel offset processing on the first image 21 to generate a second image (ie, a reference image used as a second viewing angle, not shown). In addition, for the sake of simplicity, the first image 21, the first depth information map 22, and the pixel offset 23 in Fig. 1 are taken as an example with an image width of 5 pixels (Pixel); in addition, for the sake of brevity See, Figure 1 only shows the depth information and pixel offset of 2 rows of pixels.

In the above method, if the first image 21 is subjected to the pixel offset process of shifting to the right, in the generated second image, a plurality of edge pixels in one or more rows of pixels located at the left edge must be represented by black dots. to fill in pixel values that did not exist originally. Similarly, it can be seen that if the first image 21 is subjected to the pixel shift processing of shifting to the left, the same situation will also occur at the right edge of the second image.

Since the left or right pixel offset is determined by the depth information, the effect of generating the second image by using the conventional technique will be affected by the depth information of the first depth information map 22 . For example, when the depth information corresponding to the edge pixels on the left and right sides of the first depth information image 22 is not a uniform value, the corresponding pixel offset 23 is also not a uniform value, so that the left edge or the right edge of the second image appears Uneven black image blocks. In this way, the stereoscopic image generated by the first image 21 and the second image may affect the viewing experience of the user.

In view of the above problems of the prior art, the present invention provides a 3D image generating method and a 3D image generating device to solve the problem of uneven black image blocks on the left edge or right edge of the prior art when generating the second image .

The stereoscopic image generating device provided by the present disclosure includes: a storage unit; a display unit including a display screen; and a processing unit connected to the storage unit and the display unit, and the processing unit obtains a first an image; the processing unit processes the first image to obtain depth data of each pixel of the first image, and is configured as a first depth information map, the first depth information map has the corresponding information of each pixel a depth information; the processing unit, based on a plurality of edges of the first depth information map, and a plurality of edge pixels within a predetermined width from the plurality of edges as objects, perform consistent processing, so that the processing The subsequent plurality of edge pixels have the same corresponding depth information to create a second depth information map; the processing unit sets the first depth information based on the depth information corresponding to each pixel of the second depth information map a pixel offset corresponding to each pixel of the 1 image; the processing unit for performing pixel offset processing on the first image to generate a second image; and the processing unit for outputting the first image and the first image 2 images are sent to the display unit to display a stereoscopic image.

In one embodiment, the plurality of edges are upper and lower edges of the first depth information map.

In one embodiment, the plurality of edges are left and right edges of the first depth information map.

In one embodiment, the predetermined width is 1 pixel.

In one embodiment, the depth information corresponding to each of the plurality of edge pixels in the second depth information map is the maximum depth of field in the first depth information map.

In one embodiment, the depth information corresponding to each of the plurality of edge pixels in the second depth information map is the minimum depth of field in the first depth information map.

In one embodiment, the depth information corresponding to each of the plurality of edge pixels in the second depth information map is a constant.

In one embodiment, the depth information corresponding to the plurality of edge pixels in the second depth information map is an arithmetic mean of the depth information corresponding to the plurality of edge pixels in the first depth information map.

In one embodiment, if the value of the depth information is larger, the corresponding pixel offset is smaller; if the value of the depth information is smaller, the corresponding pixel offset is larger.

The method for generating a stereoscopic image provided by the present disclosure includes: obtaining a first image, and processing the first image to obtain depth data of each pixel of the first image, and configuring a first depth information map; 1. The depth information map has a depth information corresponding to each pixel point; with a plurality of edges of the first depth information map as a reference, and a plurality of edge pixels within a predetermined width from the plurality of edges are used as objects. , carry out the uniform processing, so that the processed edge pixels have the same corresponding depth information to establish a second depth information map; based on the depth information corresponding to each pixel of the second depth information map, to set a pixel offset corresponding to each pixel of the first image; perform pixel offset processing on the first image to generate a second image; and output the first image and the second image for display A stereoscopic image.

According to the present invention, since the pixel offset of the first image is set according to the second depth information map, the second image generated by the present invention will not have uneven black image blocks on the left or right edges. .

The above and other objects and advantages of the present invention will become more apparent upon reference to the detailed description hereinafter described in conjunction with the accompanying drawings of the new model.

FIG. 2A is a schematic functional block diagram of the stereoscopic image generating apparatus of the present invention, and FIG. 2B is a hardware structure diagram of the stereoscopic image generating apparatus of the present invention.

In FIG. 2A , the stereoscopic image generating apparatus 1 at least includes: a storage unit 11 , a display unit 12 , and a processing unit 13 . FIG. 2B shows an example in which the stereoscopic image generating apparatus 1 is a notebook computer. However, the notebook computer is only an exemplary example, and the stereoscopic image generating device 1 can also be a desktop computer, a tablet computer, a smart phone, a head-mounted display, a server, a portable electronic device, or other similar computing devices. Power electronics...etc.

The storage unit 11 can be, for example, a random access memory (Random Access Memory, RAM), a read only memory (Read Only Memory, ROM), a flash memory, an erasable programmable read only memory (Erasable memory). Non-volatile memory or volatile semiconductor memory such as Programmable Random Access Memory, EPROM), Electrically Erasable Programmable Random Access Memory (EEPROM), etc.

In addition, the storage unit 11 may also be a magnetic disk, a floppy disk, an optical disk, a CD, a compact disk, or a Digital Versatile Disc (DVD).

In other words, the storage unit 11 can store any of the “first image”, “second image”, “first depth information map”, “second depth information map”, and “pixel offset” mentioned in this specification. One or a combination thereof, as well as all parameters, formulas, algorithms, program codes, etc. used by the processing unit 13 described later when performing processing.

The display unit 12 can be, for example, an output device having a display screen, such as a stereoscopic imaging display, a system-integrated panel, a light-emitting diode display, a touch screen, and the like. In other words, the display unit 12 can display the “first image”, “second image”, “first depth information map”, “second depth information map”, and “pixels” mentioned in this specification according to the needs of the user. Offset", etc., or any combination thereof is displayed on the display screen. In some embodiments, the stereoscopic image generating apparatus 1 may not include a display unit, which can output the generated stereoscopic image to an external display unit.

The processing unit 13 is connected to the storage unit 11 and the display unit 12 , and performs one-way or two-way interaction with the storage unit 11 and the display unit 12 . The processing unit 13 uses parameters, formulas, algorithms, program codes, etc. stored in the storage unit 11 to execute various processes described later. In addition, the processing unit 13 may be implemented by hardware, software, or a combination of hardware and software.

The processing unit 13 can be composed of a single circuit, a composite circuit, a programmable processor, a parallel programmable processor, a graphics processor, an application specific integrated circuit (ASIC), a field programmable gate array (Field Programmable) Gate Array, FPGA), or a combination thereof, to implement the specific functions and processes mentioned in this specification.

In addition, the processing unit 13 can also realize the specific functions and processes mentioned in this specification by reading and executing the program codes stored in the storage unit 11 . In other words, the processing unit 13 can be used to implement the "stereoscopic image generation method" mentioned in this specification.

FIG. 3 illustrates an embodiment of the operation mode of the stereoscopic image generating apparatus 1 of the present invention.

As shown in FIG. 3 , the processing unit 13 acquires the first image 31 from the storage unit 11 . The processing unit 13 uses the depth estimation model stored in the storage unit 11, such as a conventional depth estimation model such as a convolutional neural network model (but not limited to this), to obtain each pixel of the first image 31 , to generate the first depth information map 32 . The first video 31 may be the same as or different from the first video 21 shown in FIG. 1 . If the first image 31 is the same as the first image 21 , after the same depth estimation process, the first depth information map 32 (corresponding to the first image 31 ) will be the same as the first depth information map 22 shown in FIG. 2 . (corresponding to the first image 21) the same or similar. In the following, for the sake of simplicity, we assume that the first image 31 and the first image 21 are the same, and the first depth information map 32 and the first depth information map 22 are also the same.

In addition, for the sake of simplicity, the first image 31, the first depth information map 32, the second depth information maps 321A to 321D, and the pixel offsets 322A to 322D in FIG. 3 are set to have an image width of 5 pixels. (Pixel) as an example. However, in one embodiment, the first image 31 may also be an image with a size of 256×256 pixels, an image with a size of 1920×1080 pixels, an image with a size of 1024×768 pixels, or an image of any other size.

Each pixel of the first depth information map 32 has corresponding depth information. In this specification, depth information is a quantifiable value, which is defined as the relative distance between the position of a pixel in 3-dimensional space and the camera. If the value of the depth information is larger, it means that the shooting position of the pixel is farther from the camera; on the contrary, if the value of the depth information is smaller, it means that the shooting position of the pixel is closer to the camera.

In fact, depth information can change the range of values for different specifications. In this manual, for the sake of simplicity, the range of depth information is defined between 0 and 10. 0 represents the minimum depth of field that can be detected by the first depth information map 32 (ie, the closest to the camera), and 10 represents the maximum depth of field that can be detected by the first depth information map 32 (ie, the farthest from the camera).

The "depth information" has a certain correspondence with the "pixel offset" described later. Specifically, in the process of generating a stereoscopic image, a smaller pixel offset is set for a pixel with larger depth information; a larger pixel offset is set for a pixel with smaller depth information. The principle of this setting is: when we place an object in front of the eyes and perform lateral translation with the same displacement, if the object is farther away from the eyes, the lateral displacement change felt by the eyes will be smaller; The closer the distance to the eyes, the greater the change in the lateral displacement felt by the eyes. Therefore, using the characteristic of "negative correlation" between "depth information and pixel offset", the generated stereoscopic image can reflect the real feeling of the human eye when viewing the stereoscopic image. And this can also be explained back in the first figure, the correlation between the depth information of each edge pixel in the first depth information map 22 and the corresponding pixel offset.

In addition, in FIG. 1, the relational expression between the depth information and the pixel offset can be expressed by a mathematical expression of “pixel offset=21−depth information×2” (depth information: 0~10). This relational formula will continue to be used in the third and fourth figures below, so as to facilitate the reader's understanding of the technical content of this specification, but is not limited to this mathematical formula.

Next, the processing unit 13 processes the depth information of the first depth information map 32 . The processing unit 13 uses a plurality of edges of the first depth information map 32 as a reference, and takes a plurality of edge pixels within a predetermined width from the plurality of edges as objects, and performs uniform processing, so that the processed plurality of edge pixels have the same depth information to create a second depth information map.

The multiple edges mentioned here can be the upper and lower edges of the image, or the left and right edges of the image. The upper edge and the lower edge mean the pixel offset processing mentioned later, which is to perform pixel offset on the first image in an upward or downward manner; while the left edge and the right edge mean the pixel offset processing mentioned later. , which is to perform pixel shift to the left or right of the first image. Since it is common to perform pixel shift to the left or right in the conventional technology, in the following description, a plurality of edge pixels of the left edge and the right edge are used as objects.

Also, from an algorithmic point of view, we can strictly define what this specification calls an "edge" in terms of 2-dimensional coordinates. Taking an image with a size of 256x256 as an example, if the pixel in the lower left corner of the image is the origin O(0,0), the rightward is the +x direction, and the upward is the +y direction, then all the pixels whose x-coordinate is 0 The line segment formed, the side of the line segment that is not adjacent to other pixels can be defined as the "left edge"; the line segment formed by all pixels with an x-coordinate of 255, the line segment is not adjacent to other pixels. , which can be defined as the "right edge"; the line segment formed by all pixels with a y-coordinate of 255, the side of the line segment that is not adjacent to other pixels can be defined as the "upper edge"; all y-coordinates are 0. The line segment composed of pixels, the side of the line segment not adjacent to other pixels, can be defined as the "lower edge".

In addition, the "predetermined width" from any edge is in units of pixels. For example, if the predetermined width from the left edge and the right edge is 2, then according to the above definition, the object of the uniform processing is all the edge pixels whose x-coordinates are 0, 1, 254, and 255. If the predetermined width from the left edge and the right edge is 1, the object of the uniform processing is all the edge pixels whose x-coordinates are 0 and 255, and so on.

In other words, the predetermined width may be any natural number. However, in order to avoid the loss of too much information in the generated stereoscopic image, the predetermined width is generally set to 1 pixel. Fig. 3 shows an embodiment of "predetermined width=1", and Fig. 4 shows an embodiment of "predetermined width=2".

Here, the "unification processing" performed by the processing unit 13 is to adjust the depth information corresponding to the plurality of edge pixels in the first depth information map 32 to the same value. In the embodiment shown in FIG. 3, the processing unit 13 can select any one of the four implementations (A) to (D) to adjust the corresponding edge pixels of the first depth information map 32. In-depth information. In the first depth information map 32, the depth information corresponding to the pixels other than the plurality of edge pixels is maintained as it is without adjustment. In this way, a part of the adjusted depth information and the other part of the unadjusted depth information can be used to create the four second depth information maps 321A to 321D described below.

Embodiment (A) In the embodiment (A) (refer to FIG. 3 (A)), the processing unit 13 may set the depth information corresponding to the 10 edge pixels in the first depth information map 32 as the first depth The maximum depth of field in the infographic 32 is the preset 10. Therefore, the depth information corresponding to the 10 edge pixels in the second depth information map 321A is all 10.

Embodiment (B) In Embodiment (B) (refer to FIG. 3 (B)), the processing unit 13 may set the depth information corresponding to the 10 edge pixels in the first depth information map 32 as the first depth The minimum depth of field in the infographic 32, which is the preset 0 previously. Therefore, the depth information corresponding to the 10 edge pixels in the second depth information map 321B are all 0.

Embodiment (C) In Embodiment (C) (refer to FIG. 3 (C)), the processing unit 13 can set the depth information corresponding to the 10 edge pixels in the first depth information map 32 to 0~10. Any constant in between, for example, set it to 9. Therefore, the depth information corresponding to the 10 edge pixels in the second depth information map 321C is all 9.

Embodiment (D) In the embodiment (D) (refer to FIG. 3 (D)), the processing unit 13 may calculate the arithmetic mean value of the depth information corresponding to the 10 edge pixels in the first depth information map 32, and set is the arithmetic mean. In Figure 3, since the two sets of edge depth information corresponding to the 10 edge pixels in the first depth information map 32 are 7, 6, 5, 4, 3 and 9, 8, 7, 6, and 5, respectively, the calculated The arithmetic mean is 6. Therefore, the depth information corresponding to the 10 edge pixels in the second depth information map 321D is all 6.

For the second depth information maps 321A to 321D generated by the processing unit 13 according to any one of the above-mentioned embodiments (A) to (D), since the depth information corresponding to the 10 edge pixels has been consistent; therefore, when processing When the unit 13 sets the pixel offsets 322A to 322D corresponding to each pixel of the first image 31 based on the depth information corresponding to each pixel of the second depth information maps 321A to 321D, the first image 31 can be guaranteed. The 10 (2 groups) edge pixels of , and their corresponding pixel offsets are all the same.

For example, if according to the embodiment (A), the depth information corresponding to the 10 edge pixels of the second depth information map 321A is all 10 (maximum depth of field), then according to "pixel offset = 21 - depth information x2" ” (depth information: 0~10), the corresponding pixel offsets 322A are all 1.

For example, if according to the embodiment (B), the depth information corresponding to the 10 edge pixels of the second depth information map 321B are all 0 (minimum depth of field), then according to "pixel offset = 21 - depth information x2" ” (depth information: 0~10), the corresponding pixel offsets 322B are all 21.

For example, if according to the embodiment (C), the depth information corresponding to the 10 edge pixels of the second depth information map 321C are all 9 (any constant between 0 and 10), then according to the "pixel offset" Quantity=21-depth information x2" (depth information: 0~10), the corresponding pixel offsets 322C are all 3.

For example, according to Embodiment (D), the depth information corresponding to the 10 edge pixels in the second depth information map 321D is the arithmetic operation of the depth information corresponding to the 10 edge pixels in the first depth information map 32 . The average value (6 in this example) is based on the mathematical formula of "pixel offset = 21-depth information x2" (depth information: 0~10), and the corresponding pixel offset 322D is 9.

From the perspective of the above-mentioned embodiments (A) to (D), in order to simplify the algorithm of the processing unit 13 as much as possible and save program resources, we can consider directly setting the depth information corresponding to the 10 edge pixels as some constant (9). In this way, when the processing unit 13 successively processes the continuous frames of the first image 31 , it can ensure that the pixel offsets set successively remain constant and do not fluctuate with time.

In addition, in order to avoid the pixel shift processing on the first image 31, if the second image generated is excessively distorted (ie, the area of the uniform black image block is too large), the processing unit 13 can also directly convert the 10 edge pixels The depth information corresponding to the point is set to the maximum depth of field (10). In this way, the corresponding pixel offset can be ensured to be the minimum value (1), so that the area of the uniform black image block in the second image is minimized.

Therefore, the processing unit 13 according to this new embodiment is not based on the pixel offset ( 23 ) corresponding to the first depth information map 32 ( 22 ), but is based on the second pixel offset ( 23 ) that has undergone uniform processing for a plurality of edge pixels. The pixel offsets 322A to 322D corresponding to the depth information maps 321A to 321D are subjected to pixel offset processing on the first image 31 to generate a second image. In this way, when the processing unit 13 displays the first image 31 and the second image of the stereoscopic image on the display unit 12, there will be no uneven black image blocks appearing at the edges of the stereoscopic image, and the visual use The user maintains a good experience, so it can achieve the desired effect of this case.

In the above description, the case where the "predetermined width" is set to 1 is described using FIG. 3 . On the other hand, in Fig. 4, the case where the "predetermined width" is set to 2 is explained. FIG. 4 illustrates another embodiment of the operation of the stereoscopic image generating apparatus of the present invention.

The difference between this embodiment and FIG. 3 is that the processing unit 13 must convert the first depth information map 42 into five edge pixels with a width of 2 from the left edge and five edges with a width of 2 from the right edge Pixels are also included in the processing object. Therefore, in the unification process shown in FIG. 4, there will be depth information corresponding to 20 edge pixels that need to be adjusted.

Embodiment (A) In the embodiment (A) (refer to FIG. 4 (A)), the processing unit 13 can set the depth information corresponding to the 20 edge pixels in the first depth information map 42 as the first depth The maximum depth of field in the infographic 42 is the preset 10. Therefore, the depth information corresponding to the 20 edge pixels in the second depth information map 421A is all 10.

Embodiment (B) In Embodiment (B) (refer to FIG. 4 (B)), the processing unit 13 may set the depth information corresponding to the 20 edge pixels in the first depth information map 42 as the first depth The minimum depth of field in the infographic 42 is the 0 preset previously. Therefore, the depth information corresponding to the 20 edge pixels in the second depth information map 421B are all 0.

Embodiment (C) In the embodiment (C) (refer to FIG. 4 (C)), the processing unit 13 can set the depth information corresponding to the 20 edge pixels in the first depth information map 42 to 0~10. Any constant in between, for example, set it to 8. Therefore, the depth information corresponding to the 20 edge pixels in the second depth information map 421C is all 8.

Embodiment (D) In the embodiment (D) (refer to FIG. 4 (D)), the processing unit 13 may calculate the arithmetic mean value of the depth information corresponding to the 20 edge pixels in the first depth information map 42, and set is the arithmetic mean. In Figure 4, since the 4 sets of edge depth information corresponding to the 20 edge pixels in the first depth information map 42 are "5, 4, 3, 2, 1", "6, 5, 4, 3, 2" respectively , "8, 7, 6, 5, 4" and "9, 8, 7, 6, 5", so the calculated arithmetic mean is 5. Therefore, the depth information corresponding to the 20 edge pixels in the second depth information map 421D is all 5.

For the second depth information maps 421A to 421D generated according to any one of the above-mentioned embodiments (A) to (D), since the depth information corresponding to the 20 (4 groups) edge pixels has been consistent, when When the processing unit 13 sets the pixel offsets 422A to 422D corresponding to each pixel of the first image 41 based on the depth information corresponding to each pixel of the second depth information map 421A to 421D, the first For the 20 edge pixels of the image 41, the corresponding pixel offsets are all the same.

For example, if according to the embodiment (A), the depth information corresponding to the 20 edge pixels of the second depth information map 421A is all 10 (maximum depth of field), then according to "pixel offset = 21 - depth information x2" ” (depth information: 0~10), the corresponding pixel offsets 422A are all 1.

For example, if according to the embodiment (B), the depth information corresponding to the 20 edge pixels of the second depth information map 421B are all 0 (minimum depth of field), then follow the "pixel offset = 21 - depth information x2" ” (depth information: 0~10), the corresponding pixel offsets 422B are all 21.

For example, if according to the embodiment (C), the depth information corresponding to the 20 edge pixels of the second depth information map 421C is all 8 (any constant between 0 and 10), then according to the "pixel offset" Quantity=21-depth information x2" (depth information: 0~10), the corresponding pixel offsets 422C are all 5.

For example, according to Embodiment (D), the depth information corresponding to the 20 edge pixels in the second depth information map 421D is the arithmetic operation of the depth information corresponding to the 20 edge pixels in the first depth information map 32 . The average value (5 in this example) is based on the mathematical formula of "pixel offset = 21-depth information x2" (depth information: 0~10), and the corresponding pixel offset 422D is all 11.

Therefore, the processing unit 13 provided in this application is not based on the pixel offsets corresponding to the first depth information map 42 , but is based on the pixels corresponding to the second depth information maps 421A to 421D whose edge pixels have undergone uniform processing. The offsets 422A to 422D are used to perform pixel offset processing on the first image 41 to generate a second image. Therefore, when the processing unit 13 displays the first image 41 and the second image of the stereoscopic image on the display unit 12, the user will not see uneven black image blocks at the edge of the stereoscopic image, and the user can use the The user can maintain a good viewing experience, so it can achieve the desired effect of this case.

In addition, as in the embodiment in Fig. 3 (the predetermined width is 1), in the embodiment in Fig. 4 (the predetermined width is 2), the depth information can also be directly set to a certain constant, or set to the maximum depth of field. , in order to achieve the additional effect mentioned in the embodiment of FIG. 3 .

It should be noted that, in the second depth information maps 321A~321D and 421A~421D, the pixels other than the plurality of edge pixels also have their corresponding depth information and are processed by the mathematical formula "pixel offset = 21- The pixel offset converted from depth information x2" (depth information: 0~10). However, as described earlier in this specification, since this case only needs to make the depth information corresponding to the edge pixels of the second depth information map consistent, therefore, other pixels other than the edge pixels of the second depth information map correspond to Whether the in-depth information is consistent or not is not within the scope of consideration in this case (conversely, inconsistency in in-depth information is the norm). Therefore, the drawings are not particularly drawn in Figs. 3 to 4, and relevant descriptions are omitted.

In addition, although this case expresses the negative correlation between the depth information and the pixel offset by the mathematical formula of "pixel offset = 21 - depth information x2" (depth information: 0~10), the pixel offset It is also not necessarily equivalent to offsetting the same number of pixels, but is only a schematic way of conveying the degree of offset by numerical values. Whether it is a linear or non-linear relationship, the parameters therein can be adjusted with reference to other conventional techniques.

To sum up, no matter whether the processing unit 13 adopts any one of the eight embodiments described in FIG. 3 to FIG. 4 , when the processing unit 13 displays the first image and the second image of the stereoscopic image on the display unit 12 Therefore, the user will not see uneven black image blocks at the edge of the stereoscopic image, and the user can maintain a good viewing experience, thus achieving the desired effect of the present solution.

The three-dimensional image generating device of the present invention and its operation method and method have been described in detail above. It should be noted that, the above-mentioned embodiments are only illustrative of the principles and effects of the present invention, and are not intended to limit the scope of the present invention. Those skilled in the art can make modifications and appropriate changes to the embodiments without departing from the technical principles and spirit of the present invention. Therefore, the scope of protection of the rights of this new model shall be subject to the scope of the following patent application.

1: Stereoscopic image generation device 11: Storage unit 12: Display unit 13: Processing unit 21: 1st image 22: The first in-depth infographic 23: Pixel offset 31: 1st image 32: 1st in-depth infographic 321A~321D: 2nd depth infographic 322A~322D: Pixel offset 41: 1st image 42: 1st in-depth infographic 421A~421D: 2nd depth infographic 422A~422D: Pixel offset

FIG. 1 illustrates a conventional stereoscopic image generation method. FIG. 2A is a schematic diagram of a functional block of the stereoscopic image generating apparatus of the present invention, and FIG. 2B is a hardware structure diagram of the stereoscopic image generating apparatus of the present invention. FIG. 3 illustrates one embodiment of the novel stereoscopic image generation method. FIG. 4 illustrates one embodiment of the novel stereoscopic image generation method.

1: Stereoscopic image generation device

11: Storage unit

12: Display unit

13: Processing unit

Claims (9)

A stereoscopic image generating device, comprising: a storage unit; a display unit including a display screen; and a processing unit connected to the storage unit and the display unit, and the processing unit obtains a first image from the storage unit; The processing unit processes the first image to obtain depth data of each pixel of the first image, and is configured as a first depth information map, and the first depth information map has a corresponding pixel of the first image. in-depth information; The processing unit, based on a plurality of edges of the first depth information map, and taking a plurality of edge pixels within a predetermined width from the plurality of edges as objects, performs uniform processing, so that the plurality of processed The edge pixels have the same corresponding depth information to establish a second depth information map; The processing unit, based on the depth information corresponding to each pixel of the second depth information map, sets a pixel offset corresponding to each pixel of the first image; The processing unit performs pixel offset processing on the first image to generate a second image; and The processing unit outputs the first image and the second image to the display unit to display a stereoscopic image. According to the stereoscopic image generating device of claim 1, Wherein, the plurality of edges are the upper edge and the lower edge of the first depth information map. According to the stereoscopic image generating device of claim 1, Wherein, the plurality of edges are the left edge and the right edge of the first depth information map. According to the stereoscopic image generating device of claim 3, Wherein, the predetermined width is 1 pixel. According to the stereoscopic image generating device of claim 4, The depth information corresponding to each of the plurality of edge pixels in the second depth information map is the maximum depth of field in the first depth information map. According to the stereoscopic image generating device of claim 4, The depth information corresponding to each of the plurality of edge pixels in the second depth information map is the minimum depth of field in the first depth information map. According to the stereoscopic image generating device of claim 4, The depth information corresponding to each of the plurality of edge pixels in the second depth information map is a constant. According to the stereoscopic image generating device of claim 4, Wherein, the depth information corresponding to the plurality of edge pixels in the second depth information map is the arithmetic mean of the depth information corresponding to the plurality of edge pixels in the first depth information map. According to the stereoscopic image generating device of any one of claims 1 to 8, Wherein, if the value of the depth information is larger, the corresponding pixel offset is smaller; Wherein, if the value of the depth information is smaller, the corresponding pixel offset is larger.

Images