Method and system for generation, storage and distribution of omni-directional object views
Image acquisition refers to the taking of digital images of multiple views of the object of interest. In the processing step, the constituent images collected in the image acquisition step are selected and further processed to form a multimedia sequence which allows for the interactive view of the object. Furthermore, during the Processing phase, the entire multimedia sequence is compressed and digitally signed to authorize it viewing. In the Storage and Caching Step, the resulting multimedia sequence is sent to a storage servers. In the Transmission and viewing step, a Viewer (individual) may request a particular multi-media sequence, for example, by selecting a particular hyperlink within a browser, which initiates the downloading, checking of authorization to view, decompression and interactive rendering of the multi-media sequence on the end-users terminal, which could be any one of a variety of devices, including a desktop PC, or a hand-held device.
1. An image processing method for identifying figure and background for the purpose of matteing and compositing, wherein two images are input, one with a foreground object and the other without, and the background areas are taken under similar scene illumination, the method comprising the steps of:
computing the per pixel gray level absolute difference in intensity or vector color magnitude image difference between corresponding pixels in the two images;
selecting those pixels locations which are above a relatively small threshold in terms of gray level difference or vector color magnitude difference to form a mask which selects only foreground object pixels locations.
2. The method of claim 1 , wherein the masks are combined via a logical “OR” operation to generate a combined foreground object selection mask.
3. The method of claim 1 , wherein features are identified and corresponded between two frames and affine transform components are used to align the two frames.
4. The method of claim 1 , wherein an alignment wizard consisting of rotational, translation and scaling visual displays and GUIs are used to guide and assist a user in rectifying a media sequence of individual captured images.
5. The method of claim 4 , wherein the media sequence is further processed to automatically crop for a maximum inscribed rectangle in the sequence, and the maximum inscribed rectangle is further centered around an indicated axis of symmetry for an object of interest.
6. The method of claim 1 , wherein a unique database primary key corresponding to an object to be acquired is generated on a home personal computer, a bar coded encoding of the unique database primary key is printed on the home personal computer, the print out is brought to a self-contained view acquisition unit vending system and the bar code scanned, to avoid the re-keying of that unique database primary key.
7. An image editing method for identifying figure and background for the purpose of matting and compositing an object having a visual pattern or texture, which is placed on a textureless rotating platform in front of a fixed camera, and the background of the object is stationary and two or more images are captured, the method comprising the steps of:
computing optical flow or time derivative for each pixel of each image in the image sequence, for yielding a flow vector having a magnitude and direction for each pixel;
computing a magnitude for each pixel if the optical flow measure is used;
threshold and label each pixel in the flow field with flow vector magnitude greater than threshold Θ; and
selecting pixels which are above a relatively small threshold in terms of optical flow magnitude, to form a mask which selects only foreground object pixels.
8. The method of claim 7 , wherein the masks are combined via a logical “OR” operation to generate a combined foreground object selection mask.
9. The method of claim 7 , wherein features are identified and corresponded between two frames and affine transform components are used to align the two frames.
10. The method of claim 7 , wherein an alignment wizard consisting of rotational, translation and scaling visual displays and GUIs are used to guide and assist a user in rectifying a media sequence of individual captured images.
11. The method of claim 10 , wherein the media sequence is further processed to automatically crop for a maximum inscribed rectangle in the sequence, and the maximum inscribed rectangle is further centered around an indicated axis of symmetry for an object of interest.
12. The method of claim 7 , wherein a unique database primary key corresponding to an object to be acquired is generated on a home personal computer, a bar coded encoding of the unique database primary key is printed on the home personal computer, the print out is brought to a self-contained view acquisition unit vending system and the bar code scanned, to avoid the re-keying of that unique database primary key.