Significance-Linked Connected Component Analysis (SLCCA)

SLCCA is an image encoding technique which was developed at the University of Missouri - Columbia.  In cmesh, we attempt to create a pseudo-image for input to the SLCCA encoder.  This pseudo-image is comprised of the quantized vertex positions.  The order that the values are sent to the SLCCA encoder is based upon a pre-order traversal of the Vertex Tree.  All of the x-values come first, followed by the y-values, followed by the z-values.  In addition, the way the x-, y- and z-values are stored in this pseudo-file can have 2 formats - normal or snake.

Normal and Snake compression refers to how the vertex positions are stored before they are sent to the SLCCA encoder.  The image sent is an NxN image at this time.  One restriction is that the "image" sent to the SLCCA encoder should be an even-valued number.   For example, suppose a mesh has 30 vertices.  Since each vertex has an x-, y-, and z-position, there are a total of 90 values.  Since the square-root of 90 is 9.486, which is not an even number, we would round up to 10, so that a total of 100 values, in a 10x10 format, is sent to the SLCCA encoder.  In Normal format, the values are stored just like one would read an English book - from left to right, while Snake format requires the values to be wound like a snake.  An example is shown below, for 30 vertices, in a 10x10 format, for both normal and snake formats.  Note that the last z-value is repeated a number of times, to make the "image"
an NxN one.

                           NORMAL FORMAT

        x1    x2    x3    x4    x5    x6    x7    x8    x9    x10
        x11  x12  x13  x14  x15  x16  x17  x18  x19  x20
        x21  x22  x23  x24  x25  x26  x27  x28  x29  x30
        y1    y2    y3    y4    y5    y6    y7    y8    y9    y10
        y11  y12  y13  y14  y15  y16  y17  y18  y19  y20
        y21  y22  y23  y24  y25  y26  y27  y28  y29  y30
        z1    z2     z3    z4    z5    z6    z7     z8    z9    z10
        z11  z12   z13  z14  z15  z16  z17  z18   z19  z20
        z21  z22   z23  z24  z25  z26  z27  z28   z29  z30
        z30  z30   z30  z30  z30  z30  z30  z30   z30  z30
 

                              SNAKE FORMAT

        x1    x2    x3    x4    x5    x6    x7    x8    x9    x10
        x20  x19  x18  x17  x16  x15  x14  x13  x12  x11
        x21  x22  x23  x24  x25  x26  x27  x28  x29  x30
        y10  y9    y8    y7    y6    y5    y4    y3    y2    y1
        y11  y12  y13  y14  y15  y16  y17  y18  y19  y20
        y30  y29  y28  y27  y26  y25  y24  y23  y22  y21
        z1    z2     z3    z4    z5    z6    z7     z8    z9    z10
        z20  z19   z18  z17  z16  z15  z14  z13   z12  z11
        z21  z22   z23  z24  z25  z26  z27  z28   z29  z30
        z30  z30   z30  z30  z30  z30  z30  z30   z30  z30
 

More experimentation of using this technique on 3D triangular meshes is forthcoming.  To see the results, simply see the RESULTS section.  If you would like more specific information about SLCCA, please click Here.
 
 
 

If you have any questions or comments, email me at scottm@meru.cecs.missouri.edu