TWI493963B - Image generating device and image adjusting method - Google Patents

Description

Image generating device and image adjusting method

The present invention relates to image processing, and more particularly to systems and methods for grouping depth images and automatically adjusting stereo images to enhance visual effects.

With the development of stereoscopic displays, the processing of stereoscopic images is becoming more and more important. In general, stereo image acquisition can be performed in several ways, such as by using a depth camera that can obtain depth information, by a dual camera that simulates human binocular vision, or by appropriate image processing from a two-dimensional image. To get a stereo image. As shown in FIG. 1A, the process of converting from a two-dimensional image to a stereo image can be roughly divided into several steps: image reduction, edge detection, line tracing, and depth assignment. ), depth image enlargement and smoothing, lateral shifting, when the depth image is created, can be combined with the original two-dimensional image to produce a stereoscopic image. The traditional algorithm for converting two-dimensional images into stereoscopic images can also establish depth images by analyzing spatial images, edge detection, and calculating vanishing points by analyzing one or more images.

As shown in Figure 1B, visual depth perception factors can be divided into physiological factors and psychological factors. In general, the depth map of a two-dimensional image converted into a stereo image is often calculated for several psychological factors. For example, psychological factors tend to think that yellow objects, large moving objects, or large objects have the shallowest depth of field. Conversely, blue objects, small moving objects, or small objects are also considered psychological factors. The depth of field is the deepest, and the material is close to being considered the same depth of field.

Furthermore, depth of field information is the key to stereoscopic image display technology. However, when the depth map is generated, only the relative relationship of each object in the image can be defined. However, traditional stereoscopic image display technology tends to focus only on how to produce correct depth of field. The way it is produced, but the depth of field information is rarely used to further process stereoscopic images.

The present invention provides an image adjustment method, including: receiving a first image and a corresponding first depth image; dividing the first depth image into at least two groups; and grouping the first depth image Applying a corresponding depth parameter to generate a second depth image; and generating an output image according to the first image and the second depth image.

The present invention further provides an image generating device, comprising: an image classifier for receiving a first image and a corresponding first depth image, and dividing the first depth image into at least two groups; an image adjuster, The method includes: applying a corresponding depth parameter to each group in the first depth image to generate a second depth image; and a stereo image viewer for generating an output according to the first image and the second depth image. image.

The present invention further provides an image adjustment method, including: receiving a first image and a corresponding first depth image; dividing the first depth image into at least two groups; according to the groups of the first depth image, The first image is correspondingly divided into the groups; an image parameter applied to the groups in the first image to generate a second image; and an output image is generated according to the second image.

The present invention further provides an image generating device, comprising: an image classifier for receiving a first image and a corresponding first depth image, and dividing the first depth image into at least two groups, wherein the image classification The first image is correspondingly divided into the groups according to the groups of the first depth image; an image adjuster is configured to apply corresponding groups to the groups in the first image respectively. An image adjustment parameter to generate a second image; and a stereo image renderer to generate an output image according to the second image.

The present invention further provides an image adjustment method, including: receiving a first image and a corresponding first depth image; dividing the first depth image into at least two groups; receiving an external control signal, wherein the external control signal is Determining a relative position of a user and a display device; when the relative position is changed, applying a depth parameter to each group in the first depth image to generate a second depth image; and according to the first The image and the second depth image generate an output image.

The present invention further provides an image generating device, comprising: an image classifier for receiving a first image and a corresponding first depth image, and dividing the first depth image into at least two groups; an image adjuster For receiving an external control signal, wherein the external control signal indicates a relative position of a user and a display device, wherein when the relative position is changed, the image adjuster correspondingly displays the first depth image Each of the groups applies a depth parameter to generate a second depth image; and a stereo image viewer is configured to generate an output image according to the first image and the second depth image.

Figure 2 is a diagram showing an image generating apparatus 200 in accordance with an embodiment of the present invention. The image generating device 200 further includes a depth analyzer 210 and a stereo image viewer 220. The depth analyzer 210 receives an input image (which may be a pure two-dimensional image or a stereo image), and generates a two-dimensional image and a corresponding depth image according to the input image. The stereo image viewer 220 receives the two-dimensional image generated by the depth analyzer 210 and the corresponding depth image to generate a corresponding stereo image. In an embodiment, the image generating device 200 further includes a depth classifier 230 and an image adjuster 240. The depth classifier 230 divides the image content of the depth image into a plurality of groups according to image features in the depth image ( Or area), wherein the content of each group includes the number of classification groups, and the pixels corresponding to each group number (or the number of groups corresponding to each pixel), and the image adjuster 240 obtains through a judgment mechanism or receives an external control signal. The number of classification groups, depth parameters, and image parameters. Before the stereo image viewer 220 draws the stereo image, the image adjuster 240 adjusts the depth image or the two-dimensional image according to the obtained depth parameter and the image parameter, so that the stereo image drawn by the stereo image renderer 220 has different Variety.

In another embodiment, the depth classifier 230 has three ways to divide the image content of the depth image into a plurality of groups. The first type of division is regional, meaning that it is directly divided into two dimensions in the depth image. For example, if the size of the image is 1920×1080, the grouping of the depth image may be three groups of the upper 1/4, the middle 1/2, and the lower 1/4 (such as the depth image 320 of FIG. 3A), or In the middle of the depth image, a region having a width of 800 and a height of 500 is divided, and the remaining portion is another region (such as the depth image 300 of FIG. 3A), or the depth image is divided according to a predetermined region position. (The depth image 310 in Fig. 3A), but the invention is not limited thereto. The second division method is depth type, which means that the depth image is analyzed, and the group corresponding to the specific grayscale value is captured. For example, the depth image is a grayscale image. In an embodiment, the depth classifier 230 can capture an area of the depth image with a grayscale value between 180 and 255, and mark it as a region, depth. The other portions of the image are further divided into regions. The depth images 330, 340, and 350 in FIG. 3B display divided regions corresponding to different grayscale value ranges, but the present invention is not limited thereto.

The third type of division is the object type, that is, the depth classifier 230 detects the depth of field content in the depth image, and uses foreground detection to divide the foreground object and the background object, or combine the front and rear objects. The plurality of depth images corresponding to the dimensional image detect the movement of the area with the same gray range to analyze the dynamic object or the repeated object, such as a dynamic large object, a dynamic small object, and a stationary object. For example, as shown in FIG. 3C, the depth classifier 230 may divide different objects into different groups according to dynamics. In the depth images 360, 370, and 380, the characters, the car, and the ground may respectively be foreground objects. As shown in FIG. 3D, the object A blocks the object B and the object C in the depth image, and the depth classifier 230 can thereby know that the object A is a foreground object. For the identification method of the object, reference may be made to the related art of image processing. The present invention uses only the method of occlusion identification and dynamic degree identification as an example, but the present invention is not limited thereto.

In another embodiment, the depth classifier 230 can further divide the group of depth images by combining the regional and depth division modes. For example, the depth classifier 230 can be in the middle of the depth image 300 in FIG. 3A. The region looks for a specific depth range (eg, a depth range between gray and gray values between 150 and 200) to further divide the group, and the depth classifier 230 may also use the intersection of the regional and depth regions. Moreover, the depth classifier 230 can further intersect the object type, the area type, and the depth type to divide the group of depth images, but the present invention is not limited thereto.

In an embodiment, after the depth classifier 230 divides the depth image into different regions, it is also known that each group divided in the depth image also has a corresponding group (or corresponding group) in the two-dimensional image. The image adjuster 240 can apply depth parameters to different groups in the depth image, and/or apply image parameters to corresponding groups of the two-dimensional images corresponding to the depth image, that is, the image adjuster 240 can be used for depth. Adjust the image and/or 2D image. For example, the depth parameter is a parameter for adjusting the contrast, brightness, gamma, sharpness, or noise reduction of the depth image, and the image parameter is to adjust the contrast, brightness, gamma, Sharpness, noise suppression, saturation, hue, R/G/B gain, R/G/B offset, or movement and zoom of a two-dimensional image, but the invention is not limited thereto. In another embodiment, the adjustment parameters of the depth image and the two-dimensional image can be independently adjusted for each region, that is, each region can have a set of independent adjustment parameters, and the image adjuster 240 can be configured according to the depth analyzer 210. The result of the analysis is either an external signal (such as a signal output by a scaler scalar, an externally sensed signal output by a user device or an environmental device (such as a light sensor), or an external signal for static image detection) to select a specific one. Or the corresponding depth parameter, image parameter or image profile. As shown in FIGS. 7A and 7B, in different embodiments, different group adjustment parameters (including depth parameters and image parameters), or all groups, may be applied to different groups in the depth image and the 2D image respectively. The groups apply the same group group adjustment parameters together, but the invention is not limited thereto.

In an embodiment, when the user views a stereoscopic movie (converted from a two-dimensional image to a stereoscopic image), the depth analyzer 210 first analyzes the depth of the image content of each of the two-dimensional images, and when a main object is detected ( For example: the car) continuously moves from right to left in the first to ten images, and the depth of field is the intermediate depth, and the depth classifier 230 detects that the movement of another object (for example, a person) is small, so The primary object is defined as the first group, the other primary object (eg, person) is defined as the second group, and the rest of the depth image is defined as the third group. In the first to ten images, the image adjuster 240 uses respective depth parameters for different regions of the depth image, for example, the gray scale range of the first group is adjusted from 120 to 160 to 70 to 140, and 10 sheets are maintained. Image; adjust the gray scale range of the second group from 0 to 40 to 20 to 40, and maintain 10 images; adjust the gray scale range of the third group from 160 to 255 to 220 to 255, and maintain 10 sheets image. The image adjuster 240 further uses respective two-dimensional image parameters for different regions in the two-dimensional image, such as increasing the saturation of the first group, changing the second group, and blurring the third group. By adjusting the depth image and the two-dimensional image, the dynamic car has a solid three-dimensional level and clear image, the background is blurred and the depth of field is deep, and the image quality remains unchanged but the depth is slightly moved backward.

In another embodiment, when the user views the three-dimensional game (non-two-dimensional image is converted into a stereo image) and pauses the image, when the depth analyzer 210 determines that the input image is a still image, the depth classifier 230 starts to The depth image is divided into regions, for example, the game protagonist (near object) is divided into a first group, and the other portion is a second group. Image adjuster 240 further applies respective depth parameters to different groups. For example, in the depth image, the gamma value of the first group is increased, and the gray level value of the second group is set to zero. In the two-dimensional image, the first group is not adjusted, and the second group is set as a gray-scale sub-image. In this way, the main character of the game in the pause can be highlighted on the screen, while the other content is at the deepest part of the screen and is a grayscale sub-image.

4A-4C are flowcharts showing an image adjustment method according to an embodiment of the present invention. Taking the image of FIG. 4A as an example, in step S400, the depth analyzer 210 receives the input image, and generates a first image (two-dimensional image) and a corresponding first depth image according to the input image, wherein the input image can be pure two-dimensional image. The image or the stereoscopic image is, for example, a two-dimensional image and a corresponding depth image, or a left-eye image and a corresponding right-eye image. The depth image can also be generated by analyzing the left-eye/right-eye image. In step S410, the depth classifier 230 divides the first depth image into at least two groups, and correspondingly divides the first image into the group according to the group of the first depth image, wherein the grouping manner is For regional, deep or object type. In step S420, the image adjuster 240 applies a depth parameter to each group in the first depth image to generate a second depth image. In step S430, the image adjuster 240 applies an image parameter to each of the groups in the first image to generate a second image. In step S440, the stereoscopic image viewer 220 generates an output image according to the first image and the second depth image, and plays the output image on a display device (not shown in FIG. 1).

4B is a flow chart showing an image adjustment method according to another embodiment of the present invention. The difference between Fig. 4B and Fig. 4A is that the flow of Fig. 4B only adjusts the depth image. Therefore, the stereo image viewer 220 is configured to generate an output image according to the first image and the second depth image in step S450.

FIG. 4C is a flow chart showing an image adjustment method according to another embodiment of the present invention. The difference between Fig. 4C and Fig. 4A is that the flow of Fig. 4C only adjusts the first image. Therefore, in step S450, the stereoscopic image viewer 220 is based on the second image to generate an output image. It should be noted that the output image in FIGS. 4A-4C can be a two-dimensional image or a stereo image, and the display device that plays the output image can receive the two-dimensional image and the stereo image.

Figure 5A is a schematic diagram showing the adjustment of the viewing position in accordance with an embodiment of the present invention. Since the content of the screen viewed by the user on the display device is a flat content, when the user views the movie directly in front of the display device, the depth classifier 230 uses the image adjustment method in the fourth embodiment to display the image. The content object is divided into groups (or regions), but when the relative position of the user and the stereoscopic image display device 200 is changed, for example, the user moves left/right or moves back and forth, or the position of the screen moves, the user views The image content to be moved also moves with the change of the relative position, and the image adjuster 240 of the present invention can further adjust the offset for each of the divided regions or groups. For example, when the user moves their position, the deeper group moves relatively shorter distances, and the shallower depth areas move relatively longer. In an embodiment, the image adjuster 240 receives an external control signal to allow the image adjuster 240 to know the relative relationship between the current user and the stereoscopic image display device 200 (eg, up, down, left and right offset, angle, distance, etc.) . In another embodiment, the external control signal can be generated by an image detecting device (not shown in FIG. 2, such as a camera), and the image detecting device can detect relative movement between the user and the screen when When the movement is greater than a predetermined distance, the image detecting device sends an external control signal to the image adjuster 240, and repeats the continuous detecting action.

In an embodiment, as shown in FIG. 5A, the depth image 520 indicates a depth image corresponding to a screen viewed by the user directly in front of the stereoscopic image display device 200, and the depth of the house is the deepest and the depth of the person is the shallowest. The depth of the car is between the house and the person. As shown by the depth images 510 and 500, when the user moves to the left, the object in the screen is seen to move relatively to the right, and the moving distance of the house and the car is shorter than the moving distance of the person. As shown by the depth images 530 and 540, when the user moves to the right, the object in the screen is seen to move relatively to the left. In another embodiment, as shown in FIG. 5B, the depth image 560 is a depth image corresponding to a picture viewed by the user at a normal distance. As shown by the depth image 550, when the user moves backward, the image adjuster 240 will narrow down each group in the depth image, and the user will see that the objects in the screen are relatively reduced. As shown by the depth image 570, when the user moves forward, the image adjuster 240 will enlarge each group in the depth image, and the user will see that the objects in the screen are relatively enlarged. In the above embodiment, the movement of the user can be regarded as moving along a normal line parallel to the screen of the stereoscopic image display device or perpendicular to the normal line, when the moving direction of the user is perpendicular to the normal of the stereoscopic image display device. The external control signal also makes corresponding adjustments to inform the image adjuster 240 to adjust the movement of the group in the depth image in the opposite direction of the user's moving direction. If the moving direction of the user is parallel to the normal of the stereoscopic image display device, when the user approaches the stereoscopic image display device, the external control signal will also make corresponding adjustments to inform the image adjuster 240 of the group in the depth image. Group zoom in. Conversely, when the user is away from the stereoscopic image display device, the external control signal will also make corresponding adjustments to inform the image adjuster 240 to reduce the group in the depth image. In the above embodiment, the user may move in the upward, downward, left, and right directions, or may be combined with the stereoscopic image display device, and the image adjuster 240 may adjust the group in the depth image. However, the invention is not limited thereto.

6A-6B are flowcharts showing an image adjustment method according to various embodiments of the present invention. Taking the image of FIG. 6A as an example, in step S600, the depth analyzer 210 receives the input image, and generates a first image (two-dimensional image) and a corresponding first depth image according to the input image, wherein the input image can be pure two-dimensional image. Image or stereo image (2D image and corresponding depth image). In step S610, the depth classifier 230 divides the first depth image into at least two groups, wherein the manner of grouping the groups may be regional, deep, or object-based. In step S620, the image adjuster 240 receives an external control signal, wherein the external control signal indicates a relative position of a user to the stereoscopic image display device displaying the first depth image. In step S630, when the relative position is changed, the image adjuster 240 correspondingly applies a depth parameter to each group in the first depth image to generate a second depth image. In step S640, the stereoscopic image viewer 220 generates an output image according to the first image and the second depth image. It should be noted that the output image generated by the stereo image viewer 220 can be a stereo image or a two-dimensional image.

The step of FIG. 6B is different from that of FIG. 6A. In step S650 of FIG. 6B, the image adjuster 240 applies an image parameter to each group in the first image to generate a second image, and the stereo image is drawn. The device 220 generates an output image according to the second image. It will be appreciated by those skilled in the art that the present invention can adjust groups in depth images and/or two-dimensional images to produce an output image (including stereoscopic or two-dimensional images), and can also be stereoscopic The image and/or the two-dimensional image is played on at least one display device (meaning that the stereo image and/or the two-dimensional image are simultaneously output), and the display device for playing the output image can receive the stereo image or the two-dimensional image. The flow of FIGS. 6A and 6B is similar to the flow of FIGS. 4A-4C, except that the 6A and 6B images receive an external control signal to adjust the first image and/or the first depth image.

The above is only the preferred embodiment of the present invention, and the scope of the invention is not limited thereto, that is, the simple equivalent changes and modifications made by the scope of the invention and the description of the invention are All remain within the scope of the invention patent. In addition, any of the objects or advantages or features of the present invention are not required to be achieved by any embodiment or application of the invention. In addition, the abstract sections and headings are only used to assist in the search of patent documents and are not intended to limit the scope of the invention.

200. . . Image generating device

210. . . Depth analyzer

220. . . Stereo imager

230. . . Depth classifier

240. . . Image adjuster

300～350, 500～570. . . Depth image

Figure 1A is a flow chart showing a conventional algorithm for converting a two-dimensional image into a stereoscopic image.

Figure 1B is a schematic diagram showing visual depth perception factors.

2 is a block diagram showing an image generating apparatus according to an embodiment of the present invention.

3A is a schematic diagram showing the area division of a depth image according to an embodiment of the present invention.

Figure 3B is a schematic diagram showing the depth division of a depth image in accordance with an embodiment of the present invention.

FIG. 3C is a schematic diagram showing object division of a depth image according to an embodiment of the present invention.

3D is a schematic diagram showing a masked object in a depth image according to an embodiment of the invention.

4A-4C are flowcharts showing an image adjustment method according to various embodiments of the present invention.

5A-5B are schematic views showing adjustment of an ornamental position according to an embodiment of the present invention.

6A-6B are flowcharts showing an image adjustment method according to various embodiments of the present invention.

7A-7B are graphs showing depth parameters and image parameters in accordance with an embodiment of the present invention.

200. . . Image generating device

210. . . Depth analyzer

220. . . Stereo imager

230. . . Depth classifier

240. . . Image adjuster

Claims (32)

An image adjustment method includes: receiving a first image and a corresponding first depth image; dividing the first depth image into at least two groups; and applying a corresponding one to each group in the first depth image a depth parameter for generating a second depth image; and generating an output image according to the first image and the second depth image, wherein the step of generating the output image further comprises: receiving an external control signal, the external control The signal indicates a moving direction of a user; when the moving direction is up, down, left, right, front or back, the group is correspondingly adjusted to move in a reverse direction along one of the moving directions, thereby generating a a third depth image; and generating the output image according to the first image and the third depth image. The image adjustment method of claim 1, wherein the step of receiving the first depth image comprises: receiving at least one input image; and generating the first image and corresponding to the first image according to the input image The first depth image of the image. The image adjustment method of claim 1, wherein the dividing the first depth image into the groups comprises: dividing the first depth image according to a predetermined area position, a predetermined depth range, or a combination thereof For these groups. The image adjustment method according to claim 1, wherein The step of dividing the first depth image into the groups includes: detecting at least one foreground object in the first depth image; and dividing the first depth image into the groups according to the foreground object. The image adjustment method of claim 2, further comprising: correspondingly dividing the first image into the groups according to the groups in the first depth image; Each of the groups applies a corresponding image parameter to generate a second image; and generates an output image according to the second image and the second depth image. An image generating device includes: an image classifier for receiving a first image and a corresponding first depth image, and dividing the first depth image into at least two groups; and an image adjuster for Each of the first depth images is respectively applied with a corresponding depth parameter to generate a second depth image, and a stereo image viewer is configured to generate an output according to the first image and the second depth image. Image, wherein the image adjuster further receives an external control signal indicating a direction of movement of a user, wherein the image adjustment is when the moving direction is up, down, left, right, front or back Correspondingly adjusting the groups in the first depth image to move in a reverse direction along one of the moving directions, thereby generating a third depth image, and the stereo image viewer is further based on the first image and the first image Three depth images to This output image is produced. The image generating device of claim 6, further comprising a depth analyzer for receiving at least one input image, and generating the first image and the corresponding image corresponding to the first image according to the input image A depth image. The image generating device of claim 6, wherein the image classifier divides the first depth image into the groups according to a predetermined area position, a predetermined depth range, or a combination thereof. The image generating device of claim 6, wherein the image classifier further detects at least one foreground object in the first depth image, and divides the first depth image into the groups according to the foreground object. group. The image generating device of the seventh aspect of the invention, wherein the image classifier further divides the first image into the groups according to the groups in the first depth image, and the image is adjusted. The image data is further applied to the groups in the first image to generate a second image, and the stereo image viewer further generates the output according to the second image and the second depth image. image. An image adjustment method includes: receiving a first image and a corresponding first depth image; dividing the first depth image into at least two groups; and according to the groups of the first depth image, the first The image is correspondingly divided into the groups; an image parameter applied to the groups in the first image to generate a second image; Receiving an external control signal indicating a moving direction of a user; when the moving direction is up, down, left, right, front or back, correspondingly adjusting the first depth image Moving the group in a reverse direction along one of the moving directions to generate a third depth image; and generating an output image according to the second image and the third depth image. The image adjustment method of claim 11, wherein the step of receiving the first depth image comprises: receiving at least one input image; and generating the first image and corresponding to the first image according to the input image The first depth image of the image. The image adjustment method of claim 11, wherein the dividing the first depth image into the groups comprises: dividing the first depth image according to a predetermined area position, a predetermined depth range, or a combination thereof For these groups. The image adjustment method of claim 11, wherein the step of dividing the first depth image into the groups comprises: detecting at least one foreground object in the first depth image; and according to the foreground object, The first depth image is divided into the groups. The image adjustment method of claim 12, further comprising: applying a corresponding depth parameter to the groups in the first depth image to generate a second depth image; and according to the first The image and the second depth image produce a stereoscopic image image. An image generating device includes: an image classifier for receiving a first image and a corresponding first depth image, and dividing the first depth image into at least two groups, wherein the image classifier is configured according to the The first image of the first image is correspondingly divided into the groups; an image adjuster is configured to apply a corresponding image adjustment parameter to the groups in the first image. And generating a second image; and a stereo image plotter, according to the second image to generate an output image, wherein the image adjuster further receives an external control signal, the external control signal indicating a direction of movement of the user The image adjuster correspondingly adjusts the group of the first depth image to move in a reverse direction along one of the moving directions when the moving direction is up, down, left, right, front, or back. The third image is generated according to the first image and the third depth image to generate the output image. The image generating device of claim 16, further comprising a depth analyzer for receiving at least one input image, and generating the first image and the corresponding image corresponding to the first image according to the input image A depth image. The image generating device of claim 16, wherein the image classifier divides the first depth image into the groups according to a predetermined area position, a predetermined depth range, or a combination thereof. The image generating device of claim 16, wherein the image classifier further detects at least one foreground object in the first depth image, and divides the first depth image into the groups according to the foreground object. group. The image generating device of claim 17, wherein the image classifier further divides the first image into the groups according to the groups in the first depth image, and the image is adjusted. The image data is further applied to the groups in the first image to generate a second image, and the stereo image viewer further generates the output according to the second image and the second depth image. image. An image adjustment method includes: receiving a first image and a corresponding first depth image; dividing the first depth image into at least two groups; receiving an external control signal, wherein the external control signal indicates a user a relative position of a display device; when the relative position is changed, applying a depth parameter to each group in the first depth image to generate a second depth image; and according to the first image and the first Two depth images that produce an output image. The image adjustment method of claim 21, wherein the dividing the first depth image into the groups comprises: dividing the first depth image according to a predetermined area position, a predetermined depth range, or a combination thereof For these groups. The image adjustment method of claim 21, wherein the step of dividing the first depth image into the groups comprises: Detecting at least one foreground object in the first depth image; and dividing the first depth image into the groups according to the foreground object. The image adjustment method of claim 21, wherein the step of correspondingly adjusting the groups in the first depth image comprises: when the external control signal indicates that the user is perpendicular to the display device When one of the normals moves in one direction, the groups are correspondingly adjusted to move in the opposite direction of the direction to generate the second depth image. The image adjustment method of claim 21, wherein the step of correspondingly adjusting the groups in the first depth image further comprises: when the external control signal indicates that the user is parallel to the When one of the normals of the display device is away from the display device, the groups are correspondingly reduced to generate the second depth image; and when the external control signal indicates that the user is parallel to the display device When the normal is close to the display device, the groups are correspondingly enlarged to generate the second depth image. The image adjustment method of claim 21, wherein the generating the output image further comprises: correspondingly dividing the first image into the groups according to the groups in the first depth image And when the relative position is changed, applying an image parameter to the groups of the first image to generate a second image, and generating the output image accordingly. An image generating device comprising: An image classifier for receiving a first image and a corresponding first depth image, and dividing the first depth image into at least two groups; and an image adjuster for receiving an external control signal, wherein The external control signal indicates a relative position of the user and a display device, wherein when the relative position is changed, the image adjuster correspondingly applies a depth parameter to each group in the first depth image to generate a a second depth image; and a stereo image viewer for generating an output image according to the first image and the second depth image. The image generating device of claim 27, wherein the image classifier divides the first depth image into the groups according to a predetermined area position, a predetermined depth range, or a combination thereof. The image generating device of claim 27, wherein the image classifier further detects at least one foreground object in the first depth image, and according to the foreground object, the first depth image is divided into the first depth image. Group. The image generating device of claim 27, wherein the image adjuster is more correspondingly adjusted when the external control signal indicates that the user moves in a direction perpendicular to a normal of the image generating device. The groups move in opposite directions in one of the directions to produce the second depth image. The image generating device of claim 27, wherein when the external control signal indicates that the user is away from the image generating device parallel to a normal line of the display device, the image adjuster correspondingly Reducing the groups to generate the second depth image, and When the external control signal indicates that the user is adjacent to the display device parallel to the normal of the display device, the image adjuster correspondingly enlarges the groups to generate the second depth image. The image generating device of claim 27, wherein the image classifier further divides the first image into the groups according to the groups in the first depth image, and when When the relative position is changed, the image adjuster applies an image parameter to the groups of the first image to generate a second image, and the stereo image renderer generates the output image according to the second image. .