Canon Kabushiki Kaisha

Radovan V Krtolica

Roger D Melen

Olagappan Manickam

Jonathan E DeVito

Methods, apparatus, and computer readable medium for compressing connected component objects ( ) of bi-level images. The compression apparatus ( ) can take various forms including apparatus for coding a stroke of an object ( ) or for coding the entirety of the object ( ), including plural strokes. The compression apparatus ( ) typically includes a referencing module ( ) for identifying at least one reference node ( ), a coding module ( ) for successively coding pixel runs ( ) such that at least one run ( ) is coded relative to the reference node ( ) and other runs ( ) are coded relative to previously coded runs, and a closing module ( ) for terminating the process. Certain forms of the apparatus operate in a horizontal or a vertical mode only, never operate in horizontal mode during two consecutive coding operations, code each run using two code-words, and/or utilize modified Huffman coding techniques. Various compression methods of the general nature described above are also disclosed.

Linear list compression

Methods and apparatus for creating a skeletal representation ( A) of a pixel image ( ) composed of connected components ( and ). The skeletal representation ( A) is obtained by dividing each connected component ( ) into a line segment having plural slices, calculating a minimal bounding rectangle (MBR) of each line segment, replacing each line segment with a thin line approximately formed by centroid pixels of the slices ( S) inside the MBR, and connecting the resulting thin lines ( and ). One of the many benefits of using the disclosed methods and apparatus is that the resulting thin lined graph ( A), i.e., the skeletal representation, is isomorphic to the original pixel image ( ).

Graph based thinning of graphical objects

Methods, apparatus and data structures for identifying, processing and interpreting bi-level connected component objects found within graphic images. The disclosure includes methods for generating lumped graph representations ( ) of the targeted objects ( ) and for using these to recognize the objects as known objects in a reference dictionary ( ). To implement this method, a bi-level connected component object ( ) is identified within an image ( ) and the object is represented as a plurality of pixel runs ( ) of a given binary value. A lumped graph representation ( ) which captures characteristic topological features of the object is created based upon the connectivity data of the pixel runs in the object. This information is then used to recognize the object ( ) by comparing the lumped graph feature vectors with feature vectors of known objects stored in the dictionary ( ). Optionally, the dictionary ( ) can be updated such that newly recognized objects can be added to the dictionary ( ) for future recognition purposes.

Recognition and clustering of connected components in bi-level images

A sample image ( ) is segmented by an image segmentation system ( ) including a size reduction unit ( ), which reduces the size of the image ( ), and, at the same time, fills small gaps between foreground pixels. Thereafter, a connected component analyzer ( ) identifies connected components and their associated minimum bounding rectangles in the reduced image ( ). Next, a target object filter ( ) searches the connected components for target objects, making use of a target object library ( ) to identify target objects characterized by such parameters as size, shape, and texture. Finally, an inverse mapper ( ) locates the bounding rectangles of the target objects in the original sample image ( ), and extracts the associated portions of the image ( ) for analysis in a conventional image classifier ( ).

Multiple size reductions for image segmentation

An image rotation system (,) allows for rotating an image between a first image position and a second image position. The image rotation system (,) includes an orientation module (,), a partition module (portions of,), a rotation module (,), and a concatenation module (portions of,). The orientation module (,) is disposed to receive the image to identify the initial image orientation. The partition module is disposed to receive the oriented image to partition the image into at least one sub-image. The rotation module (,) is disposed to receive each sub-image of the partitioned image to rotate the pixels of each sub-image. The concatenation module is disposed to receive each rotated sub-image to concatenate the rotated sub-images to produce the image in the final image position. A method for rotating images in a limited memory environment is also disclosed.

Image rotation system

A sample image (,) is recognized by normalizing (,) the size of a sample image (,) to the size of a referent images (,); and determining (,) a set of candidate images (,) from a set of referent images (,), wherein each of the candidate images (,) is within an acceptable distance from a different binarization (,) of the sample image (,). A system (,) for image recognition includes a scanning device (,), a normalization unit (,), a distance calculation unit (,), a classification unit (,), a disambiguation unit (,), and a display device (,).

Classification-driven thresholding of a normalized grayscale image

A printed or handwritten character image is recognized by training (,) a plurality of lookup tables with a set of known referent characters; obtaining (,) a bitmap of a character image to be recognized; mapping (,) the bitmap onto a standardized character matrix; partitioning (,) the matrix into a plurality of sub-matrices; (,) determining a set of candidates for each sub-matrix; and selecting (,) a preferred candidate from among the set of candidates responsive to at least one pre-defined selection criterion. The invention is implemented by means of a scanner (,), a character mapper (,), a matrix partitioner (,), a candidate set builder (,), and a character selector (,).

Image recognition through localized interpretation

An image scaling system that enlarges images during JPEG encoding (400) and reduces images during JPEG decoding (500). The image scaling system uses N-point forward and inverse one-dimensionial scaled discrete cosine transforms (DCTs), where N is selected from among 1, 2, 3, 4, and 6. When encoding a source image (110) to an enlarged JPEG image (422), the system partitions the source image into N.times.N blocks. Each N.times.N block is transformed using the N-point scaled DCT (410). The system modifies quantization tables (418) to account for the scale of the transform and the increase in size of the image and quantizes (412) the blocks of N.times.N scaled cosine coefficients using the modified quantization tables. The resulting N.times.N blocks are enlarged (414) to 8.times.8 blocks by padding each block with coefficients having values of zero. When decoding a JPEG image (510) to a reduced output image (518), the system retrieves 8.times.8 blocks of quantized cosine coefficients from the JPEG image. The system reduces (512) each block to an N.times.N block of quantized cosine coefficients. The system modifies the quantization tables (520) retrieved from the JPEG image to account for the scale factor and the decrease in size of the image, then quantizes (514) the N.times.N blocks to produce N.times.N blocks of scaled cosine coefficients. The system performs an N-point scaled inverse DCT (516) on each block of scaled cosine coefficients. The results of the inverse DCT form the reduced output image (518). The forward and inverse DCTs are performed using efficient processes that require relatively few calculations to achieve the desired result.

Fast scaling of JPEG images

Document distribution system 10 receives user record documents 10R containing record data 10D for distribution and storage throughout a user network 10N. Controller 10C controls the operation of the document distribution system. A transmittal sheet 10T carrying transmittal symbol accompanies each record document for directing the distribution. Document scanner 10S has a document port 10P for receiving the user documents and transmittal sheets. The document scanner is responsive to the controller for scanning the transmittal symbols on the transmittal sheet and the record data on the documents. The scanner provides an electronic pixel image of the record data for distribution and storage throughout the user network. The scanner also provides a pixel image of the transmittal symbols for directing the distribution and storage. Controller display 12 is responsive to the controller for displaying operational information about the document distribution system to the user. User response device 14 is connected to the controller for permitting the user to initiate and operate the document distribution system.

Growing community of inventors

Los Gatos, CA, United States of America

Radovan V Krtolica