Image analysis – Image enhancement or restoration – Image filter
Reexamination Certificate
2000-05-19
2004-05-04
Johns, Andrew W. (Department: 2621)
Image analysis
Image enhancement or restoration
Image filter
C382S280000
Reexamination Certificate
active
06731819
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATIONS
This invention is based on and claims the benefit of priority from the Japanese Patent Applications No. 11-141491, filed May 21, 1999; No. 11-305560, filed Oct. 27, 1999; and No. 2000-006731, filed Jan. 14, 2000, the entire contents of which are incorporated herein by reference.
BACKGROUND OF THE INVENTION
This invention relates to an optical information processing apparatus, and more particularly to a compact, general-purpose optical information processing apparatus capable of various types of filtering and image processing.
When various types of information, including images and signals, are recognized or classified, the degree of similarity of those to the comparison reference are generally calculated and then they are recognized or classified.
A combination of a matched filter (MSF) and a correlating unit or a joint transform correlating unit (JTC) has been used as means for calculating the degree of similarity.
While those methods have a sufficient performance in recognizing and classifying known, simple pieces of information cut off from the background, they have to process even pieces of information contributing less to recognition and classification, when directly processing images or signals with complex features. This is one factor which causes errors.
In addition, they respond sensitively even to a slight transformation, rotation, enlargement, or reduction, which often produces errors.
As for filters that allow transformation, noise, or the like, tremendous research effort has been directed toward filters using coordinate transformation, such as Fourier Mellin (FM) transformation, Synthetic Discriminate Function (SDF) filters, Circular Harmonic Expansion (CHE) filters, and other filters of these types.
As for the processing of complex objects, many investigations have been recently made of the technique for, instead of directly processing various types of information, such as images or signals, transforming them into features contributing much to recognition or classification, object by object in pre-processing, and recognizing them on a neural network or the like, using the transformed features.
In the pre-processing, correlation values obtained from the aforementioned MSF or JTC, edge images obtained through convolution by a Laplacian filter, feature extracted images obtained by wavelet transformation or Gabor transformation, Fourier spectrum images, or the like have been used.
In general, when two-dimensional images are processed, serial calculations on an ordinary computer are not practical, because it takes an extremely long time. It is therefor clear that an optical method capable of two-dimensional batch processing is superior in time.
Although having such superiority, the optical image processing method has hardly been put to practical use.
One reason is that there is no general-purpose optical information processing apparatus which enables the user or developer to execute various types of processing (optical information processing).
Although each of the various types of filters researched and developed until now has been excellent, they represent only one part of component technology in the field of practical industrial applications. They are not so convenient as electronic image processing boards commercially available at present and have often not been used in seeking the solution to the actual problems.
FIG. 37
shows the optical information processing apparatus disclosed in Jpn. Pat. Appln. KOKAI Publication No. 8-129149 form a different point of view.
With the optical information processing apparatus, the image to be processed is displayed on a first liquid-crystal display
2
. Then, the image is read via a collimator lens
4
by the laser light emitted from a semiconductor laser
3
. A first lens
5
performs Fourier transform optically.
Then, the image subjected to Fourier transform is produced on a second liquid-crystal display
6
. Then, a filtering function form a memory
7
is displayed on the second liquid-crystal display
6
, thereby filtering the image. After the filtered image is subjected to inverse Fourier transform using a second lens
8
, a photodetector
9
can process the image.
The problem of the image processing apparatus the present invention tries to solve will be explained by reference to
FIG. 24
schematically showing the optical information processing apparatus of FIG.
37
.
The image displayed on an image display section
1101
is read by the read light generated by a read light generator section
1102
.
A Fourier transform optical system
1103
obtains the Fourier transform of light that read the image and forms its Fourier transform image in the back focal plane of the Fourier transform optical system
1103
.
The filtering of the Fourier transform image is done by displaying a filtering function on a filtering section
1104
located in the vicinity of the back focal plane of the Fourier transform optical system
1103
.
An inverse Fourier transform optical system
1105
subjects the filtered light to inverse Fourier transform. The result is obtained by a filtering image acquiring section
1106
.
In such an image processing apparatus, the spatial optical modulator, or the image display section
1101
, has been only as large as a filtering liquid-crystal display with a small number of pixels and a small display area.
In recent years, however, as the liquid-crystal apparatuses and DMD (digital micromirror apparatuses) have included more and more pixels, this provides a wider selection of spatial optical modulators, which has satisfied the need for an increasing capacity of images to be processed.
At the same time, the entire computing apparatus has been required to be more compact.
To make the computing apparatus more compact, it is necessary to shorten the focal length f of the Fourier transform optical system
1103
.
As shown in
FIG. 24
, however, when the size W
i
of the input image is made larger and the focal length f of the Fourier transform optical system is made shorter, the angle 2&thgr;
i
of the light entering the filtering section
1104
becomes greater.
For simplicity sake,
FIG. 24
shows only the DC component which has no frequency component in the frequency components included in the image displayed on the image display section
1101
.
On the other hand, the light entering the filtering section
1104
further produces diffracted light through the pixel structure of the spatial optical modulator constituting the filtering section
1104
.
The light passed through the filtering section
1104
spreads at a specific angle in the direction opposite to the direction of incidence. The diffracted rays each spread in a similar manner.
Consequently, the light of diffraction of 0th order (light of 0th order) from the object observed overlaps with the light of the +first-order or −first-order diffraction (the +first-order light or −first-order light), which makes it impossible to acquire the properly filtered image.
The another problem of the image processing apparatus the present invention tries to solve will be explained by reference to
FIG. 25
schematically showing the optical information processing apparatus of FIG.
37
.
To simplify explanation,
FIG. 37
shows a cross section of the image processing apparatus taken along the optical axis and only the rays of light corresponding to the DC component whose spatial frequency component is zero in the spatial frequency components included in the image displayed on an image display section
2001
.
In the image processing apparatus of
FIG. 37
, the image displayed on the image display section
2001
is read by the read light generated at a read light generator section
2002
.
Then, the light reading the image is inputted to a Fourier transform optical system
2003
, which forms the Fourier transform image of the input image on its back focal plane.
A filtering section
2004
is placed in the vicinity of the back focal plane on which the Fourier transform image is formed and filters the Fourier transform image.
T
Fukushima Ikutoshi
Hashimoto Takeshi
Namiki Mitsuru
Frishauf Holtz Goodman & Chick P.C.
Johns Andrew W.
Nakhjavan Shervin
Olympus Optical Co,. Ltd.
LandOfFree
Optical information processing apparatus capable of various... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Optical information processing apparatus capable of various..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Optical information processing apparatus capable of various... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3221907