Compositing is a process for inserting a foreground object into a new background scene. The object is generally photographed against a colored backing, typically blue or green. The new background scene is photographed or generated independent of the photographed object. Generally, the object and new background scenes are scanned or otherwise processed to produce high-resolution RGB (or other color) signals for computer-assisted image compositing. In the compositing process, the colored backing surrounding the object is replaced with the new background scene so that the object and the new background scene appear to be a single scene or image.

Significant advances in the field of compositing are taught by Vlahos et al. in U. S. Patent Nos. 5,515,109; 4,409,611; 4,344,085; 4,100,569; 4,007,487; and 3,595,987. Each of these references generally disclose the state-of-the-art at the time of their filings, and continue to build upon one another to advance image compositing technology. These references are primarily directed to compositing in the motion picture industry, which requires lower resolution output than required for photography and other still picture compositing applications.

The disclosures of the references are incorporated herein by reference.

A basic principle for a typical compositing system is to develop a keying or switching signal based on the presence or absence of a backing color. As noted, the backing color is generally blue or green, and the key signal is typically used to switch from the background scene to a foreground scene when the foreground image signal leaves the blue backing and enters the object area, and vice versa. Basic still image compositing works in a similar fashion.

Pure switching techniques tend to leave behind effects at the edge of the object and the new background scene. To compensate for these effects, techniques were developed to hide these"hard"edge effects of switching by purposely blurring the edges between the foreground and the background. While softening the edge effects, detail is

often lost in the boundary area, especially where individual strands of hair, netting or lace represent the boundary area. Such effects may be acceptable for motion video, but are not acceptable for still photography and images.

These switching techniques are also unable to faithfully reproduce a full linear range of semi-transparent objects. Since these switching techniques either show the foreground or the background scene through a semi-transparent subject, the object cannot be properly reproduced on the new background.

Another drawback to existing compositing techniques is the inability to remove reflections of the old background appearing on the object. These reflections have proven difficult to remove and make the placement of the object in the new background appear unnatural. Furthermore, in addition to not being able to provide high-quality, photographic images, existing systems generally require very specific background colors and lack flexibility to remove backgrounds of any color. For example, using blue as a key background, makes it difficult to deal with human subjects have blue eyes or wearing blue clothing, such as blue jeans. Using a green background provides similar difficulties for persons wearing green clothing or objects containing related colors.

Applicants are not aware of any existing system that provides unlimited flexibility in selecting a background color without substantial reconfiguration of the compositing system.

Thus, there is a need for a compositing system capable of keying on backgrounds of varying colors while providing high-resolution, photographic quality images without requiring tremendous processing resources. There is also a need for a compositing system capable of removing reflections from a key background appearing on the object to be composited and replacing those reflections with those appearing to be from the new background.

Summary of the Invention The present invention provides a solution to those needs outlined above by providing a very efficient compositing system and process capable of providing high- resolution, photographic quality images wherein the composite image of the object with the new background appears natural. In the present invention, each pixel in an image block is processed to determine the presence and amount of a select key hue. If no key hue is present in the pixel, the pixel color value for the image block is used for the corresponding pixel in the composite image. If the converse is true, wherein the image

block pixel is primarily the key hue, the corresponding pixel of a new background is used for the composite image.

In areas on or near the border between the foreground object and the old background in the image block, the amount of key hue will vary from near 0% of a hue value, where small amounts of the key hue are perceptible, to 100% of the hue value, where the hue value is the predominant hue present. For those pixels having a percentage of the key hue, a portion of a key color value is removed from the color value of the image block pixel and replaced with a portion of the color value of the corresponding new background pixel to form the composite image color value. The color value being replaced or exchanged is the entire color value and not simply the key- related hue value.

Although the portions replaced do not have to be directly proportional to one another or to the percentage of the key hue present in the image block pixel, preferably each image block pixel having the key hue will have a percentage of the key color value removed from its color value and replaced with a corresponding percentage of the background pixel color value. Such replacement causes the object to blend with the new background in a way analogous to the way the object blended with the old background to provide a natural-appearing composite wherein colors from the new background appear to reflect off the object in a fashion similar to that from the old background.

The present invention does not require analysis of the surrounding pixels and color values to determine the amount of new background to bring into any pixel in the composite image or the amount of the old image block to remove. The ability to composite images without analyzing surrounding pixels significantly decreases the amount of time and processing power necessary to composite high-resolution images from that normally required with existing techniques.

In essence, the present invention provides a composite pixel color value based upon the presence and amount of a select key hue appearing in a corresponding image block pixel. As the amount of the key hue increases, the amount of the new background color value used in the composite image pixel is increased, and the key color value in the image block is decreased proportionally. The result is a highly-efficient compositing technique providing natural looking composite images.

These and other aspects of the present invention will become apparent to those skilled in the art after reading the following description of the preferred embodiments when considered with the drawings.

Brief Description of the Drawings FIGURE 1 is a block diagram of a compositing system and image retrieval system according to one embodiment of the present invention.

FIGURE 2 is a flow diagram of a preferred compositing process according to the present invention.

FIGURE 3 is a representation of selecting and placing an object with a new background according to a preferred embodiment.

FIGURES 4 and 5 are a representation of the steps of compositing the object selected with the new background according to the present invention.

FIGURE 6 is a graph representing pixel-by-pixel color value replacement of the image block with the new background according to the present invention.

FIGURE 7 is a circular representation of hue values upon which a key hue, acceptable key hue ranges, and replacement ranges are defined according to the present invention.

FIGURES 8 and 9 are a detailed flow diagram of a preferred compositing process according to the present invention.

FIGURE 10 is a flow diagram outlining field photography and central processing using compositing techniques according to the present invention.

FIGURE 11 is a flow diagram of a process for providing and maintaining a product photo database using compositing techniques according to the present invention.

Detailed Description of the Preferred Embodiments In the following description, like reference characters designate like or corresponding parts throughout the several views. Referring now to the drawings in general, and Figure 1 in particular, it should be understood that the illustrations are for the purpose of describing preferred embodiments of the present invention and are not intended to limit the invention thereto.

As best seen in Figure 1, an image retrieval and compositing system is shown having a camera or like imaging device 10, which is associated with image storage and transfer devices or media 12. Typically, the storage and transfer devices or media 12

provide images from camera 10 to a site image processor 14 and/or a main image processor 20.

When images are taken at remote locations to be processed at a central location, the camera 10 and site image processor 14 may cooperate with a user input device 16, such as a mouse or keyboard and a display 18. In such an embodiment, images may be processed to varying degrees to provide customers with immediate feedback. For example, an image of a person or an object may be taken in front of a background that is made up of a select key color. Such key color backgrounds are referred to herein as the "old background."The site image processor 14 may retrieve the image and one or more stored backgrounds from the image storage and transfer device 12 and process the corresponding images to provide an actual or simulated composite of the customer and new background. The composite image may be displayed via display 18 and manipulated accordingly using the user input device 16. The processing power of the site image processor 14 may vary from using conventional, rudimentary compositing techniques or may actually allow real time compositing according to the present invention and as described in detail below. In a particularly rudimentary setup, the processor 14, input device 16, and display 18 may be omitted, but they are preferred to enable the photographer to assure himself or herself that the captured image will be available for processing.

In certain embodiments, the image data, and perhaps selected backgrounds, are transferred to the image processor 20. The image processor 20 has access to database 22 where various additional backgrounds, images, and/or composite images may be retrieved and stored. Image manipulation and compositing is preferably performed at the main processor 20 wherein the image may be manipulated manually using the user input device 26 and viewed on display 24. Finally, a printer 28, preferably capable of providing photographic quality images is associated with the image processor 20 for providing high-quality output. Those skilled in the art will quickly recognize that the compositing techniques taught below may be implemented on virtually any control system and may operate with or without manual interaction with object images and backgrounds. Further, the images and backgrounds may be provided from any number of sources.

According to the preferred embodiment of the present invention, compositing is performed according to the flow diagram of Figure 2 and the corresponding

representations of image manipulation in Figures 3,4, and 5. The process begins (block 100) wherein an object is selected for compositing (block 102). The object is clipped or otherwise removed according to any available technique from its existing surroundings, which is referred to as the old background (block 104). The object is clipped by selecting its pixels from the old background, and omitting the old background. A new background is selected for the given object (block 106).

Preferably, a user will be able to view the selected object and the new background concurrently and moveably overlay the selected object on the new background (block 108). Being able to move the selected object about the new background allows experimentation with placement and provides an approximation of the final composite. For example, if the new background includes a set of stairs, an image of a person may be placed at the bottom or top of the stairs or anywhere else in the background to provide the desired visual effect.

Once the object is moved to a desired location in the new background, the object may be placed and sized as necessary (block 110). Sizing may be necessary to ensure the object is placed in the background according to proper scale. Up to this point in the preferred embodiment, selecting and clipping the object from the old background and placing the clipped object in the new background may be accomplished using any given technique sufficient to provide the user with a simulation of the object and new background combination, in addition to providing the flexibility to move the object throughout the background to determine proper placement. Notably, the novel compositing technique discussed below is not necessary for these preliminary operations.

Figure 3 provides a graphic representation of the steps described in the flow chart of Figure 2 up to this point. As can be seen, a clipped image 202 is removed from an original image 200 containing the object against an old background which is represented by the closely-spaced, diagonal lines. The clipped image is placed on a new background 204 represented by widely-spaced, diagonal lines, moved and located as desired by the user. This can be accomplished with known"drag and drop"techniques. Image 204 demonstrates the ability to move and scale the clipped object as desired. The clipped image may be scaled up or down depending on the desired application. Once the clipped object is moved, scaled and placed, an approximate visual representation of a composite image 206 is provided to the user.

Referring now to Figure 4 and again to the flow diagram of Figure 2, an image block 210 encompassing the select object is defined from the original image as shown in image 208 (block 112). The image block 210 will include the selected object and part of the old background. It is important that the image block 210 include a margin or penumbra that is naturally occurring and on which pixels at the edge of the object include some of the hue of the old background. The image block may be a select pattern, such as a block or circle, sized to encompass the object. Preferably, the image block is processed to substantially track the edge or periphery of the object. In such an embodiment, most pixels having the old background hue or color will be removed from the image block until one or a few pixels having the old background form a silhouette or edge around the object border. Trimming the image block down to the perimeter of the object further reduces processing time and power for the subsequent compositing process. Generally, the"trimmed"object block can be stored or otherwise processed to determine the object edge.

Next the image block 210 is associated, aligned, and scaled (if necessary) with the new background as shown at 212 (block 114). Notably, the image block 210 is associated with a new background corresponding to the visual representation of the composite image 206 shown in Figure 3. The system software identifies the pixels overlaid in the composite image 206 and selects the desired pixels from the image 208 and locates them on the image 212.

Referring now to Figure 5 in conjunction with Figure 2, the image block 210 is compared with a corresponding background block 216 (which can be a portion of block 212) having the same dimensions as the image block 210. Preferably, a key color corresponding to the old background remaining in the image block 210 is selected (block 116). The key color may be manually or automatically selected using any number of known techniques. The key color is preferably selected according to 1) a predefined definition of color or hue values or ranges; or 2) color or hue values or ranges for the old background in the image block, original image or a reference image. Preferably, a user will simply identify a portion of the image block 210 containing the old background using a mouse, keyboard, or other input device to select a key color. Once the key color is selected or determined, the present invention provides a pixel-by-pixel analysis throughout the image block 210 with reference to the corresponding background block

216 (block 118). The result is a composite image 220 as shown in Figure 5 wherein the process ends (block 120).

If the resolution or number of pixels in the image block differs from the corresponding background block 216, a composite or average value within the one block 210,216 may be computed for a corresponding pixel in the other block. For example, if the corresponding background block 216 has four times the number of pixels as the image block 214, a four-pixel set in the background block 216 will correspond to one pixel in the image block 214. The pixel set of the background block may be averaged or otherwise processed to provide a corresponding color or hue value. The converse is true when the image block is of higher resolution than the background block.

Compositing Details The compositing technique of the present invention is now described in detail.

During compositing, each pixel in the image block 210 is processed to determine the presence and amount of a select key hue. If no key hue is present in the pixel, the pixel color value for the image block 210 is used for the corresponding pixel in the composite image. If the converse is true, wherein the image block pixel is primarily the key hue, the corresponding pixel of the background is used for the composite image. In areas on or near the border between the foreground object and the old background in the image block, the amount of the key hue will vary from near 0% of the hue value, where small amounts of the key hue are perceptible, to 100% of the hue value, where the key hue is the predominant hue present. The gradation generally occurs with the amount of key hue increasing, moving outward from the object to the old background in pixel-by-pixel steps and is an artifact of the photographic process.

For those pixels having a percentage of the key hue, a portion of the key color value is removed from the image block pixel and replaced with a portion of the color value of the corresponding new background pixel for the composite image. The color value being replaced or exchanged is the entire color value of the corresponding pixels and not simply the key or related hue value. Although the portions replaced do not have to be directly proportional to one another or to the percentage of the key hue present in the image block pixel, for the sake of an exemplary embodiment, each image block pixel having the key hue will have a like percentage of the key color value removed therefrom and replaced with a corresponding percentage of the background pixel color value. Such

replacement causes the penumbra or margin of the object to blend with the new background in a way analogous to the way the object blended with the old background, to provide a more natural-looking composite wherein colors from the new background appear to reflect off the object similar to those from the old background.

As seen with reference to the pixels represented in the left side of Figure 6, when the key hue is at 100%, the image block pixel color value is completely replaced with the color value for the corresponding pixel in the new background to form the composite image. When the key hue value is not the predominant hue present in the image block pixel, a portion of the key color value is removed from the image block pixel color value and replaced or exchanged with a like portion of the color value for the new background.

As seen in Figure 6, the amount of replacement decreases as the amount of key hue decreases. As the key hue decreases pixel-by-pixel in the penumbra, an adjacent hue will begin to predominate. The adjacent hue is the hue within the image block, and typically comes from the foreground object.

Importantly, the present compositing method does not require analysis of surrounding pixels and color values to determine the amount of new background to bring into any pixel in the composite image or the amount of the old image block to remove.

The ability to composite images without analyzing surrounding pixels significantly decreases the amount of time and processing power necessary to composite high- resolution images from that normally required with existing techniques.

Furthermore, applicants'technique provides the unique ability to compensate for reflections of the old background onto the image of the object. It is often the case where backgrounds of a select key hue, typically blue or green, are reflected onto the object being photographed. In such cases, blue or green shadows appear on the object and are difficult, if not impossible, to remove. Also, the present invention actually replaces the shadows and reflections caused by the key color and hue with color values of the new background to provide a natural and lifelike blending of the object within the new background. For example, if the key color and hue were a select green, which was reflected in different degrees and levels around and about the object, compositing using the techniques of the present invention will blend the color of the new background about and around the object at substantially the same degrees and intensity of the original key hue.

Additionally, with conventional compositing methods, any transparent portions of the object through which the old background is visible causes further problems in compositing with new backgrounds. For example, photographing a human subject would often result in having the key color appearing through glasses and between strands of hair. The present invention provides an effective and efficient method of removing key hues and colors reflected on the object, appearing within the object, and surrounding the object without time consuming and rudimentary clipping and switching techniques.

Figure 7 shows a 360 degree hue chart wherein all hues of the visible spectrum are given values between 0 and 360. Typically, red is designated 0 or 360, green is designated 120, and blue is designated 240. All hues in the spectrum fall at some point on the circle. The hue chart of Figure 7 indicates an arbitrary key hue, generally referred to with reference numeral 250, selected from any one of the hue values represented by the chart. Note the hue value for key hue 250 is approximately 90 for the given example.

An important aspect of the present invention is the ability to key on any hue value and not merely select shades of blue or green, as required in prior compositing systems.

Furthermore, the key hue may be changed quickly and efficiently for any given background or object. Such ability provides tremendous flexibility and a wide variety of compositing applications, including amusement and entertainment embodiments. In short, any color and any image may be used as a key and replaced with any other color or image. Furthermore, multiple colors may be replaced with multiple backgrounds on the same image.

The circular chart of Figure 7 is divided into five sections according to the preferred embodiment of the present invention. Each of these sections is generally located by reference to the key hue 250 and represents a range for which certain compositing processes are provided.

Since the key hue 250 is infinitesimally small, and any key color in a key background will generally vary in hue, an acceptable range along the hue continuum is defined. In the preferred embodiment, the acceptable key hue range 252 extends between a lower hue value H (BL) and an upper hue value H (BH) centered about the key hue 250. Thus, any hue value falling between H (BL) and H (BH) is treated as if it is equivalent to the key hue 250, and when detected in any pixel during compositing, the image block color value will be fully replaced or exchanged with the color value for the corresponding pixel in the new background.

On either side of the acceptable key hue range 252 lies corresponding ranges 254.

When hue values within these ranges 254 are detected for a pixel, a corresponding percentage of the key color value is removed and replaced with a percentage of the color value found in the image block. Preferably, this percentage is based on the distance between the hue value found in the pixel in the image block and the key hue 250, the acceptable key hue range limit H (BL) or H (BH), or any point therebetween. The replacement ranges 254 are defined as having hue values between H (BL) and H (L) and H (BH) and H (H), respectively. The remaining range 256 represents the hue values that are deemed not to contain the key hue.

It is important to note that the key hue 250 is the only point in the continuum that is truly 100% the desired key hue without any hue content from surrounding hues. With this model, the key hue 250 affects and blends with surrounding hues in a decreasing continuum. The outside limits of the replacement range 254 (H (L), H (H)) represent selected points along the continuum where the percentage of the key hue is negligible and preferably imperceptible. Parameters defining the key hue and the surrounding ranges 252,254 are variable in hue and range. This ability to adjust the ranges allows for correction of background imperfections, artifacts, variations in lighting, background colors, and key colors in general. In the preferred embodiment, these values can be selected by an operator and stored in parameter files for ready re-use.

Attention is now directed to the flow chart of Figures 8 and 9 wherein a preferred compositing process is described in detail. The process begins (block 300) wherein a pixel or pixel composite is selected (block 302). If necessary, a corresponding pixel in the image block is converted to an HSV (hue, saturation and value) or HSI (hue, saturation and intensity) color value (block 304). Any color space, including RGB, CYM, etc., may be used to define color value, but hue values are used to ultimately control the compositing process. In the preferred embodiment, color values are defined and manipulated in the RGB domain.

Once an HSV color value is calculated, the H, S and/or V values are compared with override limits or ranges for any one or a combination of the values. For example, even if a hue value within the range 252 of an acceptable key hue is present, if the saturation and/or intensity are at levels indicating the color is likely not part of the background to be removed (block 306), the pixel color value for the image block is used for the composite pixel color value (block 308). If this is the case, compositing for the

selected pixel is complete. If the selected pixel is the last pixel in the image block for compositing (block 312), the process ends (block 314). If the pixel is not the last pixel in the image block, the next pixel is selected (block 302) and the process continues.

Assuming the selected pixel is not within any of the override limits (block 306), the pixel hue, referenced as HUE (I), is processed to see if the hue value falls between H (BH) and H (H), the upper replacement range 254, depicted in Figure 7. If it is within the replacement range, the process will calculate a percentage (more properly, a decimal between 0 and 1) of the key hue in the image block pixel (H%) (block 320). Preferably, the percentage is determined based on a range of hue values about the key; a value within or defining the range, such as H (BH), or the exact hue value of the key color. In the preferred embodiment, the percentage is calculated according to the following formula when HUE (I) falls in the upper replacement range 254: H% = (H (H)-HUE (I))/ (H (H)-H (BH)) Once this percentage is calculated, an equivalent portion of the key color value is removed from the image block pixel color value and replaced with a corresponding portion of the new background pixel or pixel composite color value. In the preferred embodiment, the RGB color value for the image block pixel is manipulated as follows: Rc = Rib-Rk * H% + Rnb * H% where Ré is the red value for the composite image; Rnb is the red value for the new background pixel; Rib is the red value for the image block pixel; and Rk is the red value of the key color value; Gc =Gib-Gk*H% +Gnb*H% where Gc is the green value for the composite image; Gnb is the green value for the new background pixel; Gib is the green value for the image block pixel; and Gk is the green value of the key color value; and Bc = Bib-Bk * H% + Bnb * H% where Bc is the blue value for the composite image; Bnb is the blue value for the new background pixel; Bib is the blue value for the image block pixel; and Bk is the blue value of the key color value (block 322). Once compositing for the pixel is complete, the process again checks to see if the selected pixel is the last pixel in the image block (block 312) and either ends compositing (block 314) or continues on with the next pixel (block 302).

If HUE (I) is not between H (BH) and H (H) (block 316), the process next checks to see if HUE (I) is between H (L) and H (BL) (block 324). If HUE (I) is between H (L) and H (BL), which indicates the hue value for the pixel is in the lower replacement range 254, the percentage of the key color value to remove from the image color value and replace with the new background color value is calculated as follows: H% = (HUE (I)-H (L))/ (H (BL)-H (L)) (block 326). Once the portion of the color value to exchange or replace is determined, color value replacement is accorded the same as that described immediately above (block 322). Notably, although the equations above calculate the replacement percentage with reference to the borders of the acceptable key hue ranges 252 (H (BL) and H (BH)), the percentages may be calculated with reference to the key hue or any value related thereto.

Also, the multiplier H% need not be a linear interpolation of the distance between the pixel hue and the defined range values, but could be determined using some other, non- linear scale. The percentage calculations are simply representative of one way of replacing a corresponding portion of the image block color value with a related portion of the new background color value based on an amount of the key hue present in the image block color value.

If the image block hue, HUE (I), is not within the replacement ranges 254, the process next checks to see if the hue falls between H (BL) and H (BH), the acceptable key hue ranges 252 (block 328). If the image block hue HUE (I) falls within the acceptable key hue ranges 252, the new background color value is used for the composite pixel color value (block 330). In this case, the process determines that the pixel contains sufficient amounts of the key hue to warrant complete replacement or exchange of the image block pixel color value with that of the new background pixel. Once the composite color value is set to the new background color value, the process continues until the last pixel in the image block is processed.

Lastly, if the image block hue value, HUE (I), is not within any of the override limits (block 306), replacement ranges 254, or within the acceptable key hue ranges 252, the resulting composite pixel color value will be that of the image block. In this situation, the process assumes there is insufficient key hue in the image block color value to warrant any modification or replacement of a portion of the image block color value (block 308). In essence, the image block color value is kept as the object color value without being affected by the new background color value. Again, the process will

continue until the last pixel in the image block is processed. Importantly, the processing efficiency of the compositing techniques of the present invention are further enhanced because compositing is necessary only throughout the image block; however, the compositing techniques may be used throughout the entire area corresponding to the composite image.

The following is an exemplary compositing routine in pseudo code: Convert RGB to HSV values; IF the saturation value is between a select range, and intensity value is between a select range, THEN IF HUE (I) is between H (BH) and H (H) THEN (HUE (I) IN UPPER REPLACEMENT RANGE) H% = (H (H)-HUE (I))/ (H (H)-H (BH)) Subtract a portion of key <BR> <BR> Rc = Rib - Rk * H% + Rnb * H% color value from image<BR> <BR> block and replace with a Gc = Gib-Gk * H% + Gnb * H% portion of new background <BR> <BR> * based on amount of hue<BR> Bc = Bib - Bk * H% + Bnb H%<BR> <BR> present in image block ELSEIF HUE (I) is between H (BL) and H (L) THEN (HUE (I) IN LOW REPLACEMENT RANGE) H% = (HUE (I)-H (L))/ (H (BL)-H (L)) Subtract a portion of key <BR> <BR> color value from image<BR> <BR> block and replace with a Gc =Gib-Gk*H% +Gnb*H% portion of new background <BR> <BR> based on amount of hue<BR> Bc = Bub - Bk * H% + Bnb * H%<BR> present in image block ELSEIF HUE (I) is between H (BH) and H (BL) THEN (HUE (I) WITHIN ACCEPTABLE KEY HUE RANGE) H% = 1; Replace color value of <BR> <BR> image block with new<BR> <BR> background Gc = Gnb ; Be Bnb ELSE (HUE (I) not present in image block)

H% = 0; (Keep image block color value) <BR> <BR> <BR> <BR> Rc = Rib;<BR> Gc = Gib; Bc = Bib END IF ELSE H% = 0 ; (Keep image block color value) END IF Field Photography Figure 10 shows one exemplary application of the compositing techniques described above for field photography. In this embodiment, images of foreground objects, such as people, are taken against key backgrounds at a variety of remote locations such as in schools and churches. In these environments, the people whose images are taken may desire to have any of a number of backgrounds in the resulting photograph. Instead of taking pictures of the subjects in front of all of the available backgrounds, pictures are taken against the key background, and desired new backgrounds are composited with the image of the subjects during subsequent processing. Thus, the photographer can focus on getting that perfect shot of the subject without concern for the background or its effect on the subject. For example, a subject may want his or her picture in front of a particular stadium, building, or landmark or within any number of selected scenes. Taking the image and subsequently applying any number of backgrounds saves time while providing tremendous flexibility.

The flow of field photography would typically begin (block 400) by obtaining or retrieving image data for a given object or objects (block 402). Customer preferences identified as order information, including background selection, size, quantity, placement within the background, etc. is entered (block 404). The image data and the order information are associated with one another and processed to provide a select file having both image and order information data for a given customer (block 406). The data may be transferred directly or remotely to a central processing site (block 408) via any number of given data transfer techniques, including electronic transfer and physical

transfer of hard drives or CD ROMs and the like. At the central processing facility, images are processed by the compositing system as discussed above (block 410) to provide the selected background, object placement, as well as providing the requested quantities and sizes. Although the process may be fully automated, manual manipulation or error correction may be necessary for certain images (block 412).

The ability of the current system to use any hue value as a key allows operators to select all or a portion of an image and key on any color within the image or portion.

Further, the ability to vary the ranges for defining the key hue, acceptable key hue ranges 252 and replacement ranges 254 allows tremendous flexibility to recognize or ignore various artifacts or areas within the image or image portion. In other words, compositing with the same or different parameters about the same or different key values may be iteratively applied throughout the image or any portion thereof to fine tune and perfect the composite image. Once the foreground and background images are processed and a composite image is obtained, the images are printed or developed (block 414) and the process ends (block 416).

Maintaining and Updating a Product Photo Database Similar to the field processing above, product databases associated with various indicia to identify the product, date of photo, etc. may be kept. The product files are stored to allow retrieval and combination with varying backgrounds. A basic process for updating and maintaining a product photo database is shown in Figure 11. The process begins (block 500) wherein image data for objects or products are obtained (block 502) from existing files, storage devices, cameras, or images. Associated product information is gathered (block 504) relating to product name, operation, project, company, size, quantity, background, and placement of the object within the background. Any number of desired parameters are acceptable and within the scope of the present invention. As noted above, the image data and information are associated with one another and processed (block 506) to provide a record for the database. The record is transferred to the accompanying directory or database (block 508) for storage and subsequent retrieval.

The record may be retrieved and processed as necessary with the associated information (block 510) to provide images necessary for product advertising and layout (block 514).

As with the field photography, manual adjustments may be made to facilitate layout requirements, fine tune compositing, and eliminate artifacts and imperfections (block 512) wherein the process ends (block 516).

Although the field photography and product layout embodiments were discussed in detail, other commercial, entertainment, and amusement applications will benefit from the present invention. In any image processing environment wherein select portions of an image are applied to another image to form a composite image, the present invention will provide an enhanced final product in a highly efficient manner.

Certain modifications and improvements will occur to those skilled in the art upon reading the foregoing description. It should be understood that all such modifications and improvements have been omitted for the sake of conciseness and readability, but are properly within the scope of the following claims.