catalog.txt

NIST Handwriting OCR Testbed

                     (search col maj) and returns the blob "stats"
 end_findblobs - deallocates memory upon completion of a findblob session.

 f_fit_param3_image2 - returns a transformed binary image by applying
                       precomputed Least Squares parameters and quickly
                       pushing pixels and conditionally filling speckle
                       noise in the resulting image (fastest).
 grp4comp - CCITT Group 4 compresses an image.

 grp4decomp - CCITT Group 4 decompresses an image.

 compute_hist - computes a horizontal or vertical spatial histogram
                from a binary bitmap.
 compute_x_hist - computes a horizontal spatial histogram from a
                  binary bitmap.
 compute_y_hist - computes a vertical spatial histogram from a
                  binary bitmap.
 hist_nruns_hori - given a list of horizontal runs, computes a histogram
                   by summing the top-n longest runs in each row.
 hist_nruns_hori2 - given a list of horizontal runs, computes a histogram
                 by summing the top-n longest runs in each row and returns
                 the sorted order of the runs.
 cut_image - returns the subimage from a binary bitmap that is
             byte aligned in origin and width.
 cut_exact_image - returns the exact subimage requested from a binary
                   bitmap.
 imagedup - takes an image of specified depth and duplicates its
            contents in a new image.
 PixPerByte - returns the pixel per byte factor given the depth of
              a pixel in an image.
 SizeFromDepth - computes the byte size of an image based on the
                 image's pixel width, pixel height, and pixel depth.
 allocate_image - allocates and zeroes (using calloc) an image of
                  specified width, height, and depth.
 mallocate_image - allocates an image of specified width, height, and
                  depth using malloc().
 find_first_max_forward - returns the location and value of the first
                          maximum searching forward through a
                          specified section of a list of integers.
 find_first_min_forward - returns the location and value of the first
                          minimum searching forward through a
                          specified section of a list of integers.
 find_first_min_backward - returns the location and value of the first
                           minimum searching backward through a
                           specified section of a list of integers.
 erode_charimage - set true pixel to zero if any of 4 neighbors is zero
                   in a character image.
 dilate_charimage - set false pixel to one if any of 4 neighbors is one
                    in a character image.
 get_south8 - return value of char image pixel 1 below of current pixel
              if defined else return (char)0
 get_north8 - return value of char image pixel 1 above of current pixel
              if defined else return (char)0
 get_east8  - return value of char image pixel 1 right of current pixel
              if defined else return (char)0
 get_west8  - return value of char image pixel 1 left  of current pixel
              if defined else return (char)0
 max_overlap_run_along_x - analyzes a system of horizontal runs and
                assigns a vector of y values accoriding to the maximum
                length run at each x posistion.
 init_markers - initializes markers to a specifiec x position to keep
                track of processing within the system of horizontal runs.
 image_pad - pads an image to a new byte-aligned width and height based
             on new width and height dimensions.
 find_pix_north - from a given point in an 8-bit pixel image, finds the
                  the first pixel upward that has the specified value.
 find_pix_south - from a given point in an 8-bit pixel image, finds the
                  the first pixel downward that has the specified value.
 find_pix_east - from a given point in an 8-bit pixel image, finds the
                 the first pixel rightward that has the specified value.
 find_pix_west - from a given point in an 8-bit pixel image, finds the
                 the first pixel leftward that has the specified value.
 find_pix_fwd_on_line - searches forward along a specified segment of a line
                 trajectory for the first pixel that has the specified value.
 find_pix_bwd_on_line - searches backward along a specified segment of a line
                 trajectory for the first pixel that has the specified value.
 n_pixels_on_line - counts the number of pixels with specified value along
                 a given line trajectory.
 sub_column_eq - checks if the designated portion of a column in an 8-bit
                 pixel image equals the specified pixel value.
 sub_row_eq - checks if the designated portion of a row in an 8-bit
              pixel image equals the specified pixel value.
 next_pix_in_region_ptr - resumes search for a pixel in a region of a binary
                      char image specified by a top and bottom trajectory using
                      pointers.
 next_pix_in_region - resumes search for a pixel in a region of a binary char
                      image specified by a top and bottom trajectory spanning the
                      entire width of the image.
 black_pixel_count - returns the total number of pixels in a binary
                     image with value equal to one.
 ReadBinaryRaster - reads the contents of a binary IHead image file
                    into an IHead structure and image memory.
 grayscale_rotate_90 - rotates a grayscale image exactly 90 degrees
                       clockwise by copying pixel rows to pixel columns.
 grayscale_rotate_n90 - rotates a grayscale image exactly -90 degrees
                 counter-clockwise by copying pixel rows to pixel columns.
 get_runs8 - extracts and returns a list of all the horizontal runs from
             the binary charimage passed.
 next_run - resumes search and extracts the next horizontal run in the
            binary charimage passed.
 add_run - adds the start point and the length of a run to the list
           of runs.
 realloc_runs - allocates (and reallocates if necessary) the list of
                runs as the list becomes full.
 segblobs8 - takes a char image and returns all the connected components in a list.

 segblobs8rlen - takes a char image and returns all the connected components and the
                 list of run lengths for the entire image.
 squashx8_8 - takes a charimage and squashes it by a factor of 8 in both
            dimensions without throwing away rows and columns (~4 X slower).
 subimage_copy - copies an 8-bit subimage to another image at a
                 specified location.
 subimage_and - uses a logical AND operator to copy an 8-bit subimage
                to another image at a specified location.
 subimage_or - uses a logical OR operator to copy an 8-bit subimage
               to another image at a specified location.
 subimage_xor - uses a logical XOR operator to copy an 8-bit subimage
                to another image at a specified location.
 subimage_add - copies an 8-bit subimage to another image adding the
                subimage pixels to the pixels at the specified location.
 subimage_sub - copies an 8-bit subimage to another image subtracting
                subimage pixels from pixels at the specified location.
 subimage_lookup - addresses a lookup table with the pixels from an
                   8-bit subimage and assignes the lookups to the
                   destination image at the specified location.
 subimage_addcon - adds a constant to the pixels in an 8-bit subimage.

 subimage_subcon - subtracts a constant from the pixels in an 8-bit
                   subimage.
 subimage_multcon - multiplies a constant to the pixels in an 8-bit
                    subimage.
 subimage_divcon - divides the pixels in an 8-bit subimage by a
                   constant.
 subimage_fmultcon - multiplies a floating point constant to the pixels
                     in an 8-bit subimage.
 subimage_fdivcon - divides the pixels in an 8-bit subimage by a
                    floating point constant.
 subimage_set - sets all the pixels in an 8-bit subimage to a specified
                value.
 subimage_zeroes - sets all the pixels in an 8-bit subimage to zero.

 subimage_ones - sets all the pixels in an 8-bit subimage to one.

 subimage_invert - inverts all the pixels in an 8-bit subimage.

 find_first_thresh_forward - finds the position and value of the first
                             integer that exceeds a threshold searching
                             forward in a specified section of a list.
 find_first_thresh_backward - finds the position and value of the first
                              integer that exceeds a threshold searching
                              backward in a specified section of a list.
 writeihdrfile - writes the contents of an IHead structure and an
                 image memory to the specified file.
 zoom8 - scales an unsigned char image based on the x & y factors
         passed.
 zoom - scales a binary bitmap up or down depending on the x & y factors
        passed.
 zoom_x - scales a binary bitmap up or down based on the x factor
          passed.
 zoom_y - scales a binary bitmap up or down based on the y factor
          passed.
 enlarge_x - expand a binary bitmap's columns by the x factor passed.

 enlarge_y - expand a binary bitmap's rows by the y factor passed.

 shrink_x - down sample a binary bitmap's columns by the x factor
            passed.
 shrink_y - down sample a binary bitmap's rows by the y factor passed.

 copy_cols - copies a single column of pixels to a range of column
             locations in a destination image.
 copy_col - copies a column of pixels to a column location in a
            destination image.
 copy_rows - copies a single row of pixels to a range of row locations
             in a destination image.
 copy_row - copies a row of pixels to a row location in a destination
            image.
 zoomN - shrinks image by logical OR of square tiles, handles 1/8 bit
         per pixel input -and- outputs, padding of outputs, prealloc
         or internal alloc of output data.
 zoomx4 - shrinks an unsigned char image down by logical OR of tiles
          of size 4x4. wrapper call to zoomN
 zoomx3 - shrinks an unsigned char image down by logical OR of tiles
          of size 3x3. wrapper call to zoomN
 zoomx2 - shrinks an unsigned char image down by logical OR of tiles
          of size 2x2. wrapper call to zoomN
 zoomx8 - shrinks an unsigned char image down by logical OR of tiles
          of size 8x8. wrapper call to zoomN
 zoomxN - shrinks an unsigned char image down by logical OR of tiles
          of size NxN. wrapper call to zoomN
 zoomaux{2,3,4,8,N} - auxialliary zooming routines called by zoomN.