BACKGROUND

[0001] Digital images and videos are represented by pixels that each has a color value. The number of colors that can be utilized to represent an individual pixel is a function of the number of bits utilized to describe the color. Increasing the number of bits used to represent colors increases the number of colors that can be represented. This is referred to as bit depth.

[0002] Images are often stored and/or transmitted at a color bit depth of eight bits per pixel per color. When viewed on displays having high resolution and high contrast, objectionable visual artifacts become noticeable as a result of the color bit depth. A common and very noticeable visual artifact is referred to as "banding" or "false contouring," which occurs when an area of an image that should be a smooth, gradual transition between colors (i.e., a gradient), is instead represented by bands of color with noticeable borders between them. These artifacts can be eliminated by storing images and video at a higher bit depth, where the color value of each pixel is represented by more bits. Many previously recorded images are, however, already stored at low bit depths, such as eight bits per pixel per color. In addition, increasing the bit depth of an image causes a corresponding increase in the space required to store the image and the amount of data that must be transmitted in order to send the image over a network.

SUMMARY

[0003] This disclosure relates in general to increasing bit depth in images and videos.

[0004] One aspect of the disclosed embodiments is a method for processing an image having a first bit depth. The method includes performing two or more iterations of a bit depth enhancement operation that increases the bit depth of the image to a second bit depth that is higher than the first bit depth. The bit depth enhancement operation includes dividing the image into a plurality of areas, performing an edge detection operation to identify one or more areas from the plurality of areas that do not contain edge features, and applying a blur to the one or more areas from the plurality of areas that do not contain edge features. In a first iteration of the of the bit depth enhancement operation, the plurality of areas includes a first number of areas, and the number of areas included in the plurality of areas decreases with each subsequent iteration of the bit depth enhancement operation. [0005] Another aspect of the disclosed embodiments is an apparatus for processing an image having a first bit depth. The apparatus includes a memory and a processor configured to execute instructions stored in the memory to perform two or more iterations of a bit depth enhancement operation that increases the bit depth of the image to a second bit depth that is higher than the first bit depth. The bit depth enhancement operation includes dividing the image into a plurality of areas, performing an edge detection operation to identify one or more areas from the plurality of areas that do not contain edge features, and applying a blur to the one or more areas from the plurality of areas that do not contain edge features. In a first iteration of the of the bit depth enhancement operation, the plurality of areas includes a first number of areas, and the number of areas included in the plurality of areas decreases with each subsequent iteration of the bit depth enhancement operation.

[0006] Another aspect of the disclosed embodiments is a non-transitory computer- readable storage device including program instructions executable by one or more processors that, when executed, cause the one or more processors to perform operations for processing an image having a first bit depth. The operations include performing two or more iterations of a bit depth enhancement operation that increases the bit depth of the image to a second bit depth that is higher than the first bit depth. The bit depth enhancement operation includes dividing the image into a plurality of areas performing an edge detection operation to identify one or more areas from the plurality of areas that do not contain edge features, and applying a blur to the one or more areas from the plurality of areas that do not contain edge features. In a first iteration of the of the bit depth enhancement operation, the plurality of areas includes a first number of areas, and the number of areas included in the plurality of areas decreases with each subsequent iteration of the bit depth enhancement operation.

[0007] In some implementations of these disclosed embodiments, the number of areas included in the plurality of areas decreases by approximately 75 percent with each subsequent iteration of the bit depth enhancement operation.

[0008] In some implementations of these disclosed embodiments, dividing the image into a plurality of areas is performed using a constant area size for the first iteration and each subsequent iteration. In other implementations, dividing the image into a plurality of areas is performed by using a first area size for the first iteration and increasing the area size for each subsequent iteration.

[0009] In some implementations of these disclosed embodiments, each area from the plurality of areas includes a plurality of pixels each having a pixel value, and the edge detection operation includes comparing a maximum pixel value in each area to a minimum pixel value in each area. In a variation of these implementations, an edge feature is identified when the difference between the maximum pixel value and the minimum pixel value in a respective area from the plurality of areas exceeds a threshold value.

[0010] In some implementations of these disclosed embodiments, each area from the plurality of areas includes a plurality of pixels each having a pixel value and applying the blur includes averaging the pixel values in each of the one or more areas that do not contain edge features.

[0011] Variations in these aspect, implementations and embodiments are described in additional detail hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The description herein makes reference to the accompanying drawings wherein like reference numerals refer to like parts throughout the several views, and wherein:

[0013] FIG. 1 is a block diagram showing an example of an environment in which a system for increasing the color bit depth of images can be implemented;

[0014] FIG. 2 is a block diagram showing an example of a hardware configuration for a receiving station;

[0015] FIG. 3 is an illustration showing an example of an input image;

[0016] FIG. 4 is a block diagram showing operation of the system for increasing the color bit depth of images;

[0017] FIG. 5 is a flowchart showing an example process for increasing the color bit depth of images;

[0018] FIG. 6 is an illustration showing a first example of dividing an image into areas and detecting edge features in the areas;

[0019] FIG. 7 is an illustration showing a second example of dividing an image into areas and detecting edge features in the areas; and

[0020] FIG. 8 is a block diagram showing an example of an image enhancement function for increasing color bit depth of images.

DETAILED DESCRIPTION

[0021] The systems and methods described herein are directed to increasing the bit depth of digital images by applying a threshold-based blur at multiple scales. By applying a blur at multiple scales, the size of the blur is automatically determined. By applying a threshold at each scale, the need for explicitly segmenting an image based on its contents is avoided. Each blurring operation and thresholding operation can be applied separately to discrete areas of the image, such as groups of pixels, to localize the effects of the blurring and to maintain edge features. Optionally, a filtering operation can be applied using an offset copy of the image to ensure smooth interpolation.

[0022] FIG. 1 is a block diagram showing an example of an environment 100 in which a system for increasing the bit depth of images can be implemented. Environment 100 includes a transmitting station 110 and a receiving station 120 that are connected by a network 130. Transmitting station 110 and the receiving station 120 can be implemented in many ways, such as in the form of one or more computing devices having a memory and a processor. Network 130 can be or include any or all of the Internet, a local area network (LAN), wide area network (WAN), virtual private network (VPN), cellular telephone network or any other means of transferring information from the transmitting station 110 to the receiving station 120.

[0023] Receiving station 120 is operable to increase the bit depth of images by processing them according to the methods taught herein. The images can then be output for display to a user at receiving station 120. As used herein, images can be individual still images such as digital photographs, or frames from a video.

[0024] In one implementation, the images that are processed by receiving station 120 are transmitted from transmitting station 110 to receiving station 120 via network 130, and are processed by receiving station 120 after they are received. For example, the images can be frames from a video stream. In response to a request from receiving station 120, transmitting station 110 can transmit the video stream to receiving station 120. In some implementations, the video stream is encoded by transmitting station 110 before

transmission. In other implementations, the video stream has been previously encoded and stored at transmitting station 110. The video stream is encoded at a first color bit depth. After receiving the video stream and encoding the video stream, receiving station 120 processes images (i.e., frames) from the video stream according to the methods taught herein in order to increase the bit depth of the images.

[0025] FIG. 2 is a block diagram of an example of a hardware configuration for receiving station 120 of FIG. 1. The same hardware configuration or a similar hardware configuration can be used to implement transmitting station 110. Receiving station 120 can include a CPU 210. CPU 210 can be a conventional central processing unit. Alternatively, CPU 210 can be any other type of device, or multiple devices, capable of manipulating or processing information now-existing or hereafter developed. Although the disclosed examples can be practiced with a single processor as shown, e.g. CPU 210, advantages in speed and efficiency can be achieved using more than one processor.

[0026] Receiving station 120 can include memory 220, such as a random access memory device (RAM). Any other suitable type of storage device can also be used as memory 220. Memory 220 can include code and data 222 that can be accessed by CPU 210 using a bus 230. Memory 220 can further include one or more application programs 224 and an operating system 226. Application programs 224 can include software components in the form of computer executable program instructions that cause CPU 210 to perform the operations and methods described here.

[0027] A storage device 240 can be optionally provided in the form of any suitable computer readable medium, such as a hard disc drive, a memory device, a flash drive, or an optical drive. One or more input devices 250, such as a keyboard, a mouse, or a gesture sensitive input device, receive user inputs and can output signals or data indicative of the user inputs to CPU 210. One or more output devices can be provided, such as a display device 260. Display device 260, such as a liquid crystal display (LCD) or a cathode -ray tube (CRT), allows output to be presented to a user, for example, in response to receiving a video signal.

[0028] Although FIG. 2 depicts CPU 210 and memory 220 of receiving station 120 as being integrated into a single unit, other configurations can be utilized. The operations of CPU 210 can be distributed across multiple machines (each machine having one or more of processors) which can be coupled directly or across a local area or other network. Memory 220 can be distributed across multiple machines such as network-based memory or memory in multiple machines. Although depicted here as a single bus, bus 230 of receiving station 120 can be composed of multiple buses. Further, storage device 240 can be directly coupled to the other components of receiving station 120 or can be accessed via a network and can comprise a single integrated unit such as a memory card or multiple units such as multiple memory cards. Receiving station 120 can thus be implemented in a wide variety of configurations.

[0029] FIG. 3 is an illustration showing an example of an input image 300 that can be processed using the systems and methods disclosed herein. Input image 300 is a digital image that is defined by pixels, as is well known in the art. Each pixel has a color value that defines the color of the respective pixel when it is output for display. Each color value can be defined by multiple component values, for example, as in well-known color spaces such as the RGB color space, the YUV color space, the HSV color space, or the HSL color space, which can represent pixels as tuples of component values. [0030] Input image 300 includes high detail areas 310 and low detail areas 320. High detail areas 310 are portions of input image 300 that include significant variations in neighboring pixel values. These variations are referred to herein as edge features. Low detail areas 320 of input image 300 include subtle variations in their color values, for example, a gradual and progressive change in color values (e.g. a gradient). For instance, high detail areas 310 might be the foreground portions of a particular image, where features in the image are close and in focus, while low detail areas 320 are background portions of the image.

[0031] Input image 300 can be stored at a low color bit depth, such as a color bit depth of eight bits per pixel per color. This can lead to the number of available colors being inadequate to represent the subtle variations of color that would naturally be present in the image. Instead, visual artifacts such as false contours 330 may be visible in the image when it is output for display, such that a viewer of the image may be able to perceive visible transitions from one color value to the next in low detail areas 320.

[0032] FIG. 4 shows operation of a system 400 for increasing the color bit depth of images. System 400 receives input image 300 as its input and generates an output image 440 as its output. System 400 can be implemented, for example, at receiving station 120 by execution of computer program instructions by receiving station 120.

[0033] Initially, input image 300 can be encoded in according to a video or still image format or any type. Input image 300 is received as an input at a decoding subsystem 410 that decodes the input image into pixel values in any suitable form, as is well known in the art. As examples, the pixel values can be in RGB format or YUV format.

[0034] The pixel values representing input image 300 are then passed to a bit depth enhancement subsystem 420. Bit depth enhancement subsystem 420 increases the bit depth of input image 300 from a first bit depth to a second bit depth that is higher than the first bit depth.

[0035] Bit depth enhancement subsystem 420 is operable to increase the color bit depth of input image 300 by applying a blur at multiple scales but without a predetermined size. Instead, the size of the blur is determined dynamically by detecting edge features in input image 300. For example, for each of several progressively larger blur sizes, the image can be divided into areas— areas that contain edge features are not blurred in order to retain the edge features, and areas that do not contain edge features are blurred. In one example, blurring is performed within an area of an image by averaging the pixel values within the area, and setting all of the pixels to the average value or modifying all of the pixel values based on the average value. Within areas lacking edge features, the blurring operations cause the pixel values of the image to change to color values that were not available at the original color bit depth of the image, thereby increasing the amount of color information in the image.

[0036] In one implementation, bit depth enhancement subsystem 420 performs two or more iterations of a bit depth enhancement operation by dividing the image into a plurality of areas, performing an edge detection operation to identify one or more areas from the plurality of areas that do not contain edge features, and applying a blur to the one or more areas from the plurality of areas that do not contain edge features. In a first iteration of the bit depth enhancement operation, the plurality of areas includes a first number of areas, and the number of areas decreases with each subsequent iteration of the bit depth enhancement operation. The number of areas can be decreased by increasing the pixel-wise size of the areas, or by maintaining a constant pixel- wise size for the areas while downscaling the image with each iteration. When multiple images are created at different scales, they are subsequently upscaled and combined, with areas in the combined image taken from the image at the largest scale at which edge features were identified in a respective area.

[0037] A filtering subsystem 430 can filter the image to remove visual artifacts, if any, that may be created by bit depth enhancement subsystem 420. Filtering subsystem 430 can perform a filtering operation subsequent to the processing performed by bit depth enhancement subsystem 420 or in parallel with the processing performed by bit depth enhancement subsystem 420 by applying filtering at each of the progressively larger blur sizes. In one example of the filtering operation applied by filtering subsystem 430, each blurring step applied by bit depth enhancement subsystem 420 is performed with respect to the a first copy of the image and a second copy of the image that has been offset by a one- pixel diagonal shift. The two copies of image are then averaged to generate output image 440. Other filtering methods can be utilized to generate output image 440. In some

implementations, filtering subsystem 430 is omitted, and output image 440 is generated by bit depth enhancement subsystem 420 without filtering.

[0038] FIG. 5 is a flowchart showing a first example of a process 500 for increasing the color bit depth of images. The operations described in connection with process 500 can be performed at one or more computers, such as at receiving station 120. When an operation is described as being performed by one or more computers, it is completed when it is performed by one computer working alone, or by multiple computers working together. The operations described in connection with process 500 can be embodied as a non-transitory computer readable storage medium including program instructions executable by one or more processors that, when executed, cause the one or more processors to perform the operations. For example, the operations described in connection with process 500 could be stored at memory 220 of receiving station 120 and be executable by CPU 210 thereof.

[0039] Operation 510 includes obtaining an input image, such as input image 300.

The input image can be obtained, for example, by accessing the image from memory, accessing the image from a storage device, or receiving the image via a network transmission such as a transmission from transmitting station 110 to receiving station 120 via network 130.

If necessary, the image can be decoded to obtain pixel values for the image.

[0040] After operation 510, a first iteration of an image processing sub-process begins, starting with operation 520 and also including operations 530, 540, 550 and 560. In the first iteration of the image processing sub-process, a blur will be applied at a first scale. In each iteration of the sub-process, the scale of the blur is changed.

[0041] In operation 520, the image is divided into a plurality of areas. The scale of the blur is controlled by dividing the image into areas in operation 520. The image is divided based on an initial scale for the first iteration, and at a changed scale for subsequent iterations, as will be explained further herein.

[0042] In operation 530, an edge detection operation is performed to identify high detail and low detail regions of the image. The edge detection operation can be applied individually to each of the areas defined in operation 520. Thus, the edge detection operation can determine, for each of the areas defined in operation 520, whether that respective area contains edge features. The edge detection operation is affected by the size of the areas defined in operation 520. In particular, the size of the areas increases in each iteration relative to the original size of the image. As the areas increase in size, the likelihood of finding an edge in each are also increases. The result, as will be explained further herein, is that the strength of the blur is increased in each iteration, but the area of the image being affected by the blur decreases.

[0043] In one example, each area from the plurality of areas includes a plurality of pixels each having a pixel value and the edge detection operation includes comparing a maximum pixel value in each area to a minimum pixel value in each area. In this example, an edge feature is identified as being present within a respective area of the image when the difference between the maximum pixel value and the minimum pixel value in a respective area from the plurality of areas exceeds a threshold value. For example, the threshold value could be one, such that edges are considered to be present if there is a difference of more than one color shade with respect to the original color bit depth of the input image. Other methods of detecting edges are known and can be utilized in operation 530. [0044] The results of the edge detection operation are stored for later use. For example, a value such as one or zero can be assigned to each area to indicate whether edge features are present in the area. In one implementation, these values are stored as an edge map in the form of a matrix of values that each represents a corresponding area of the image, with positions in the matrix corresponding spatially to locations in the image. Also, the edge map can be stored for each iteration of the image processing sub-process, such that the edge maps from each iteration of the image processing sub-process are available for use subsequent to conclusion of the image processing sub-process.

[0045] In operation 540 a blur is applied. The scale of the blur is defined by the areas defined in operation 520. In one example, each area from the plurality of areas defined at operation 520 includes a plurality of pixels each having a pixel value, and applying the blur at operation 540 is performed by determining an average value based on all of the pixel values in the respective area and setting all of the pixel values in the area to the average pixel value.

[0046] The blur is applied to the low detail regions of the image, such as one or more areas that do not contain edge features. The blur can be applied based on the edge features detected at operation 530 by blurring all of the areas of the image that do not contain edge features. In one implementation, only those areas of the image that do not contain edge features are blurred, and the areas of the image that contain edge features are not blurred. In another implementation, the entire image is blurred, but applied only to low detail regions of the image by subsequent use of the edge map for its respective iteration of the image enhancement sub-process during generation of the output image.

[0047] In operation 550, a determination is made as to whether a condition is met.

Meeting the condition indicates that a suitable number of iterations of the image enhancement sub-process have been performed, and that no more iterations need be completed. In one example, the condition is met when a predetermined number of iterations have been performed. In another example, the number of iterations to be performed is calculated. As an example, the number of iterations could be determined based on the size of the image, an initial size for the areas, and the number of times that the areas can be scaled up based on a scaling factor. In an example where the areas are 2 pixels wide by 2 pixels high in the first iteration and the size of the areas is doubled with each iteration relative to the initial size of the image, the number of iterations could be calculated as the base 2 logarithm of a largest dimension of the image.

[0048] If the condition of operation 550 is not met, another iteration of the image processing sub-process will be performed. Process 500 proceeds to operation 560, where the scale at which the blur will be applied is changed. As one example, this can be done by setting a larger size (in pixels) for each of the areas that will be defined at operation 520, such as by doubling the width and height of each area. If the width and height of each area are doubled, the number of areas in the next iteration of the image processing sub-process will decrease by approximately 75 percent relative to the previous iteration of the image processing sub-process, with the decrease being approximate as a result of rounding errors. As another example, this can be done by scaling down the image and using the same pixel size for the areas. In this example the image is downsampled by half in the width and height dimensions while the pixel- wise size of the areas remains the same, with the result being that each area represents a larger portion of the image than in the previous iteration. This will result in a 75 percent decrease in the number of areas relative to the previous iteration of the image processing sub-process, with the decrease being approximate as a result of rounding errors. The downsampled copy of the image would subsequently be upsampled when generating the final output image. This type of implementation reduces the complexity of the computations to be performed in each iteration and implicitly blurs the image by downscaling and later upscaling the image. In both of these examples, the plurality of areas defined at operation 520 is largest in a first iteration of the image processing sub-process, and the number of areas decreases with each subsequent iteration.

[0049] Subsequent to operation 560, the process returns to operation 520 where an additional iteration of the image processing sub-process commences.

[0050] If the condition of operation 550 is met, no further iterations of the image processing sub-process are performed, and process 500 advances to operation 570. In operation 570, an output image is generated, such as output image 440. The output image combines the results of the image processing performed in each iteration of the of image processing sub-process. If any of the images are downsampled images, they are upsampled to the original size of the input image prior to being combined.

[0051] The images resulting from each iteration are combined such that portions of the image that are identified in each iteration as including edges are retained and incorporated in the final image and other portions of the image are blurred. Thus, features from the image are retained from the iteration with the smallest area size (relative to the original size of the image) for which an edge feature is present. In this manner, edge features are retained in the output image while applying blur to the remainder of the image at the largest size possible that will not result in a loss of edge features. [0052] By way of example, output image 440 can be generated by starting with a blurred copy of the image from the final iteration, at which the largest scale blur was applied. For any areas of the image that were determined to have edge features, those areas are incorporated from the previous iteration, at which a lesser degree of blur was applied. Thus, edge features are retained from the iteration in which their respective area was first identified as including edge features, and at smallest area size where they were first identified as including edge features.

[0053] FIG. 6 is an illustration showing a first example of dividing an image into areas and detecting edge features in the areas. In this example, the image is downsampled for each iteration of the edge detection and blur operations, and dividing the image into a plurality of areas is performed using a constant area size for the first iteration and each subsequent iteration.

[0054] A first generation image 610 is divided into a first number of areas 612 that each includes a certain number of pixels. For example, each area can be a two-pixel by two- pixel square. Other sizes and geometries can be used for the areas. The first generation image includes one or more high detail features 614. As a result of the edge detection operation, such as a threshold applied to the minimum and maximum pixel values in each of areas 612, some of areas 612 are identified as edge features areas 616 (indicated by hatching). When the final output image is created, edge feature areas 616 from first generation image 610 will be included, while other portions of the image will be represented by portions of later- generation images. First generation image 610 is downsampled to define a second generation image 620, which is divided into areas having the same pixel size as in area 612 of the first generation, but representing a larger area of the original image. Thus, there is a smaller number of areas 622 in the second generation than the number of areas 612 in the first generation. Edge feature areas 626 are identified in second generation image 620 in the manner previously described. In the output image, the image information from second generation image 620 will be used for those portions of the image that were identified as having edge features in second generation image 620, but not in first generation image 610. Second generation image 620 is downsampled to define a third generation image 630, which is divided into areas 632 of a further reduced number but at the same pixel size. Edge feature areas 636 are identified in the manner previously described, and can be incorporated in the final image, which is generated as previously described.

[0055] FIG. 7 is an illustration showing a second example of dividing an image into areas and detecting edge features in the areas. This example is distinguishable from the example of FIG. 6 in that the images are not downsampled. Iinstead, the pixel width and pixel height of the areas is increased. Thus, in the example of FIG. 7, dividing the image into a plurality of areas is performed by using a first area size for the first iteration and increasing the area size for each subsequent iteration.

[0056] A first generation image 710 includes a first number of areas 712, and one or more high detail features 714. Areas 712 that include high detail features 714 are identified as edge feature areas 716. A second generation image 720 is the same width and height in pixels are first generation image 710, but with a blur applied within each of areas 712 as previously described. Second generation image 720 includes a smaller number of areas 722, each having a greater pixel size (e.g. width and height measured in pixels) as compared to areas 712, and edge detection is applied to identify edge feature areas 726. A third generation image 730 is then generated based on the blurred second generation image 720, divided into areas 732 of a smaller number but greater size, and edge detection is again applied to identify edge feature areas 736. The final image is created as previously described.

[0057] FIG. 8 is a block diagram showing an example of a function 800 for increasing color bit depth of images that performs iterations of edge detection and blurring in a recursive operation. Function 800 receives three images as input, which are denoted as input 802, max_in 804 and min_in 806. Function 800 generates an output image denoted as output image 808.

[0058] In a first iteration of function 800 when function 800 is called normally, input

802, max_in 804, and min_in 806 are copies of a single input image. In subsequent iterations of function 800 when it is called recursively, input 802, max_in 804, and min_in 806 will each include distinct information.

[0059] When function 800 is entered, a stacking operation 810 is performed with respect to each of input 802, max_in 804, and min_in 806. With respect to each of these images, stacking operation 810 makes a copy of the image and places it adjacent to the original image, but with the copy having a one pixel diagonal shift. By doing this, two copies of each of input 802, max_in 804, and min_in 806 will be processed concurrently, to allow filtering to subsequently be performed by simply averaging the original with the copy.

[0060] Next, a downscaling operation 820 is applied to each of input 802, max_in

804, and min_in 806. Downscaling operation 820 operates on groups of pixels in each image, thereby dividing each image into areas as previously described. For example, downscaling operation 820 can divide each image into groups of pixels and calculate a single value for each group, such that the resulting downsampled version of the image will include one pixel corresponding to each group. Downscaling operation 820 can downscale input 802 by outputting an average value for each pixel group, downsample max_in 804 by outputting the maximum pixel value for each pixel group, and downsample min_in 806 by outputting the minimum pixel value for each pixel group.

[0061] The downsampled images for max_in 804 and min_in 806 are provided to an edge detection operation 830, which in this case compares the difference of max_in 804 and min_in 806 and compares that difference to a threshold to define an edge map that is upsampled in an upscaling operation 840 and passed as an input to an image combining operation 870.

[0062] The downsampled images for input 802, max_in 804 and min_in 806 are also provided to a recursive function call 850, which causes function 800 to be called recursively. Recursive function calls 850 will continue for each iteration of function 800 until a condition is met, as described with respect to operation 550 of process 500. For example, a desired number of iterations can be set as a variable that is passed to function 800 when it is called, and this variable can be decremented for each generation until the desired number of iterations have been performed, at which point function 800 will be concluded without an additional instance of recursive function call 850.

[0063] The result of recursive function call 850 is an output image 852 from the recursive instance of the function, which is generated in the same manner as described with respect to output image 808 as a result of completion of function 800. Output image 852 is passed to an upscaling operation 860, with the upscaled output image from recursive function call 850 being provided to image combining operation 870.

[0064] Image combining operation 870 combines the stacked version of input 802 with the upscaled output image from recursive function call 850. These images are combined using the edge map. For each area of the image, the corresponding area from input 802 is kept if an edge feature is present (e.g. the difference of max_in 804 and min_in 806 is not less than a threshold, which can be represented as a zero in the edge map), and the corresponding area from the output image from recursive function call 850 is kept if an edge feature is not present (e.g. the difference of max_in 804 and min_in 806 is less than a threshold, which can be represented as a one in the edge map).

[0065] The combined image created by image combining operation 870 is provided to an unstacking operation 880, which separates the two stacked copies that were created in stacking operation 810 into two separate images, which are averaged in an averaging operation 890, thereby filtering the image to remove artifacts and resulting in output image 808. As a result of the blurring performed by the combination of downscaling operations 820 and the subsequent upscaling operation 860, the pixels values in output image 808 can be expressed at a higher color bit depth than used for input 802.

[0066] Stacking operations 810, unstacking operation 880 and averaging operation

890, in combination, perform a filtering operation in parallel with bit depth enhancement, as explained with respect to filtering subsystem 430. Function 800 can be used without filtering, in which case, stacking operations 810, unstacking operation 880 and averaging operation 890 would be omitted.

[0067] In the description above, each of stacking operations 810 results in a single pixel-shifted copy being stacked with the original image, and unstacking operation 880 results in two images that are then averaged at averaging operation 890. The number of copies that are stacked with the original image can be modified, and filtering operations that involve more than one copy of the image are explicitly contemplated. For example, three copies could be stacked with the original image, with the original image being unshifted, a first copy being shifted horizontally by one pixel, a second copy being shifted vertically by one pixel, and a third copy being shifted diagonally by one pixel.

[0068] The words "example" or "exemplary" are used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as "example" or "exemplary" is not necessarily to be construed as preferred or advantageous over other aspects or designs. Rather, use of the words "example" or "exemplary" is intended to present concepts in a concrete fashion. As used in this application, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or". That is, unless specified otherwise, or clear from context, "X includes A or B" is intended to mean any of the natural inclusive permutations. That is, if X includes A; X includes B; or X includes both A and B, then "X includes A or B" is satisfied under any of the foregoing instances. In addition, the articles "a" and "an" as used in this application and the appended claims should generally be construed to mean "one or more" unless specified otherwise or clear from context to be directed to a singular form. Moreover, use of the term "an implementation" or "one implementation" throughout is not intended to mean the same embodiment or implementation unless described as such.

[0069] Implementations of transmitting station 110 and/or receiving station 120 (and the algorithms, methods, instructions, etc., stored thereon and/or executed thereby) can be realized in hardware, software, or any combination thereof. The hardware can include, for example, computers, intellectual property (IP) cores, application- specific integrated circuits (ASICs), programmable logic arrays, optical processors, programmable logic controllers, microcode, microcontrollers, servers, microprocessors, digital signal processors or any other suitable circuit. In the claims, the term "processor" should be understood as encompassing any of the foregoing hardware, either singly or in combination. The terms "signal" and "data" are used interchangeably. Further, portions of transmitting station 110 and receiving station 120 do not necessarily have to be implemented in the same manner.

[0070] Transmitting station 110 and/or receiving station 120 can be implemented using a general purpose computer or general purpose processor with a computer program that, when executed, carries out any of the respective methods, algorithms and/or instructions described herein. In addition or alternatively, for example, a special purpose

computer/processor can be utilized which can contain other hardware for carrying out any of the methods, algorithms, or instructions described herein.

[0071] Further, all or a portion of implementations of the present invention can take the form of a computer program product accessible from, for example, a tangible computer- usable or computer-readable medium. A computer-usable or computer-readable medium can be any device that can, for example, tangibly contain, store, communicate, or transport the program for use by or in connection with any processor. The medium can be, for example, an electronic, magnetic, optical, electromagnetic, or a semiconductor device. Other suitable mediums are also available.

[0072] The above-described embodiments, implementations and aspects have been described in order to allow easy understanding of the present invention and do not limit the present invention. On the contrary, the invention is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims, which scope is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structure as is permitted under the law.