Image analysis – Applications
Reexamination Certificate
2000-05-01
2003-04-15
Tran, Phuoc (Department: 2621)
Image analysis
Applications
C382S285000
Reexamination Certificate
active
06549639
ABSTRACT:
FIELD OF INVENTION
The present invention relates to imaging of human body parts. More particularly, the present invention relates to imaging the edges of a body part or an object, such as a human foot including the footprint, arch and instep using a patterned background.
BACKGROUND OF THE INVENTION
Conventional measurements of 3-D objects have used a laser scanners or area image sensors that scans the area of the object to be imaged. The sensors may be CMOS or charged coupled device (CCD) sensors, such as used in a digital camera. The scanners may be a single line CCD scanner or a PC scanner. Generally, the overall dimensions of an object are measured by imaging the edges of the object. The periphery of the object is constructed when imaging object edges that are contrasted from a different colored background. The edge of the object is interpreted using software algorithms that discerns transitions from the edge of the object from the space behind the object. The space behind the object has to have some degree of color from white to black different from the object so that the data input to an image processing algorithm can indicate the edge of the object. The edges of an object are often blurred or indiscernible from the background as a result of shadowing or the use of similar colors between the background the object. When the object being measured has any edges of a color similar to the background, the algorithm fails to accurately detect the edges, and hence the algorithm inaccurately determines the edges of the object. When the object is multicolored, it is often difficult to accurately discern the edges of the object as the background will merge wit the edges of the object. The inability to accurately detect edges leads to inaccurate sizing of three-dimensional (3D) objects. When contour imaging a 3D object using imaged edges, the scanning or sensing means circumscribes the 3D outline of the object by moving the sensor around the object to image a plurality of edges around the object. The 3D image of the object then constructed will be made up of the continuous outside dimensions of the object and will accurately display the outermost dimensional contour of the object.
Other methods of imaging 3D objects include laser imaging methods that measured the depth of laser projected light beams as the beams are reflected back to a detector using beam deformation and position changes to image depth. Typically, the detector used with the laser detects the position of the laser beam over the entire surface of the object as the laser beam is moved. The laser imaging method often fails due to color, texture and reflection of the imaging laser beam. In both scanning and laser methods, the software algorithms are complex because the imaging process requires scanning and detecting areas of objects having light intensity such that the detector means may not be able to definitive discern the edge of the object. To detect the edges, prediction algorithms are used to fill in areas of the object that are imprecisely detected. Typically, the software algorithm is complex and slow and inherently unreliable. These and other disadvantages are solved or reduced using the present invention.
SUMMARY OF THE INVENTION
An object of the invention is to provide a system and method that indirectly images objects by contrasting the object with a predetermined background pattern.
An object of the invention is to provide a system and method that indirectly images objects by contrasting the object with a predetermined background checkerboard grid pattern of alternating white and black colored areas.
Another object of the invention is provide a system and method for cross-referencing human body and body parts dimensions to manufacture apparel sizes for accurately ordering apparel form manufacturers.
Another object of the invention is to provide a means of capturing both the weighted and non-weighted variations of feet to enable the imaging of arch height and type, and the spread of the feet under weight.
The invention is directed to an imaging system and method that indirectly measures an object using a background pattern. An object to be measured is placed in front of a predetermined background grid. Imaging means image the background pattern that is interrupted by the object placed in front of the background pattern. The system and method images the background pattern. When the ordered regularity of the background pattern is interrupted by the object, the edge is accurately determined by counting the number of alternating areas, such as black and white areas, from a known border of the background pattern or from other predetermined fixed reference locations, targets or purposeful irregularities within the background pattern. Software algorithms then determine the overall measurement of the object where the background pattern has been interrupted to indirectly measure the periphery of the object. A plurality of images taken from different perspectives provide accurate 2D peripheral measurements that are combined to provide a 3D measurement of the object being measured.
The imaging means may include an array of light detecting cells such as charged coupled devices (CCD) or complementary metal oxide silicon (CMOS) area sensors. These sensors are commonly used in a digital cameras having appropriate lenses that focus the image to impinge the image over an image area onto the area sensors. The image in digital form can be stored in a processing system such as a personal computer. A computer processing system is used to process the stored image during an imaging process. A plurality of images may be stored for respective different angle positions of the imaging means. Preferably, for each image recorded, the imaging process counts repetitive marks to reconstruct the visible area of the background pattern. The repetitive marks are preferably blocks in rows and columns of alternating black and white blocks of a preferred checkerboard background pattern. The repetitive marks are counted starting from a known edge position usually known as a reference target.
The imaging process processes the stored image by determined the edge of the object relative to interruption in the regularity of the background pattern. An expected error of one or two blocks, that is, the background pattern image tolerance error, can occur depending on the color of the object being imaged contrasted from the color of the background pattern. Each background pattern, area, mark or block, may be as small as is detectable by the sensor means. In the case of a checkerboard pattern imaging a human foot, for example, the block dimension may be only one millimeter and well within sensor and focusing lens capabilities. Conventional edge average smoothing processes maybe be performed during computer image processing to obtain an average edge contour line depicting the edge of the object accurate to plus or minus one block dimension. Each mark or block represents a fixed dimension. Counting the number of marks or blocks from a plurality of reference targets or from fixed patterned borders will enable by imaging processing, a method for determining the dimensions of the object being measured within one mark or block dimension.
The background pattern only needs predetermining contrasting marks for image recognition and processing. Alternating checkerboard blocks is the simplest to make and use as the preferred background pattern. However, there are several other types of background patterns that may be used. For example, Fresnel patterns may be used where each block deflects the light to appear like lighted and unlighted blocks to the sensor means. For another example, a sublighted platform where black blocks are printed on a clear or translucent bottom material so that light emanates upward through the light translucent blocks. The sublighted platform is advantageous because it eliminates top lighting shadows. Other patterns such as circles, rectangles, triangle, hexagons, among many other, could be used as well, for different applications and accuracy c
Genovation Inc.
Reid Derrick Michael
Tran Phuoc
LandOfFree
Body part imaging system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Body part imaging system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Body part imaging system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3055334