Structure-guided automatic learning for image feature...

Image analysis – Image enhancement or restoration – Edge or contour enhancement

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C382S155000

Reexamination Certificate

active

06507675

ABSTRACT:

TECHNICAL FIELD
This invention relates to image processing methods to automatically learn image feature enhancement processing sequence and parameters.
BACKGROUND OF THE INVENTION
Many computer vision applications require the enhancement and detection of image features for use in classifiers for detecting and classifying objects of interest or to prepare images for measurement operations. Application domain knowledge is available in most of the computer vision applications. The application domain knowledge can often be expressed as structures of image features such as edges, lines and regions, patterning or texture, color, straightness, or fit information (i.e. associated part information). The structures include spatial configurations and relationships of object features such as shape (rectangular, circular and directional, etc.), size, intensity distribution, parallelism, co-linearity, adjacency, intersection angles, etc. The structure information can be well defined in industrial applications such as semiconductor manufacturing, electronic assembly or machine part inspections. In machine part inspections, most of the work-pieces have Computer Aided Design (CAD) data available that specifies its components as entities (LINE, POINT, 3DFACE, 3DPOLYLINE, 3DVERTEX, LINE, POINT, 3DFACE, 3DPOLYLINE, 3DVERTEX, etc.) and blocks of entities. In biomedical or scientific applications, structure information can often be loosely defined. For example, a cell nucleus is round and different cell shapes differentiate different types of biological materials such as blood cells or chromosomes.
A structure-guided processing invention was disclosed in U.S. patent application Ser. No. 09/738846 entitled, “Structure-guided Image Processing and Image Feature Enhancement” by Shih-Jong J. Lee, filed Dec. 15, 2000 that provides a sub-pixel, high performance image feature extraction and enhancement through a structure-guided image processing method which is incorporated in its entirety herein. This US Patent Application teaches how application domain structure information is encoded into the processing sequences and parameters for structure-guided extraction and enhancement of features of interest and removal of noisy and irrelevant information.
In video inspection and measurement applications, objects of different designs are inspected. Image processing sequence and parameters are defined by a designer during the set-up phase for an inspection target. Image processing sequence and parameters have to be updated whenever target designs are changed.
Prior Art
Prior art relies on human experts using a trial and error method to perform the update process. This task is manageable in simple applications where image features have high contrast and low noise and appropriately trained personnel are available. However, in many practical applications, feature contrast and signal to noise ratio are low and the processing results are significantly impacted by even a small change of the processing sequence and/or parameters. For these applications, the human update process is impractical since it is time consuming and the results are not consistent. The resulting system performance depends on the skill and experience level of the human expert. Unfortunately, the availability of skilled human experts is limited and they are often expensive. If the objects to be inspected/measured change frequently, this significantly impacts the productivity, feasibility, and utility of an image inspection/measurement system.
OBJECTS AND ADVANTAGES
It is an object of this invention to reduce the cost to set up or update video measurement or inspection system image processing sequences and/or parameters.
It is an object of this invention to automatically generate an image processing recipe with consistent results.
It is an object of this invention to allow a relatively low skill operator to set-up, maintain, and effectively use an image processing system.
It is an object of this invention to create new applications for video or image measurement and inspection systems. The new applications would adapt easily to frequently changing target specifications.
It is an object of the invention to improve the performance of machine vision systems by easing their effective application.
SUMMARY OF THE INVENTION
One of the limiting factors in the application of image processing technology is the need for skilled personnel to design algorithms that are purpose specific. Current methods produce inconsistent results and are costly and time consuming. This invention provides a method to automatically learn a recipe for image feature enhancement processing sequence and parameters. The learning method of this invention uses data from the application domain structure information and a target detection specification input to produce the image processing recipe and an enhancement goodness measure. An application module uses the image feature enhancement processing sequence and parameter recipe, to process an input image wherein a detection target and application domain structure are indicated using calipers. The application module produces a feature enhanced image output. An enhancement goodness measure is used to evaluate and choose between alternatives in the learning method.


REFERENCES:
patent: 5295197 (1994-03-01), Takenaga et al.
patent: 5315700 (1994-05-01), Johnston et al.
patent: 5329596 (1994-07-01), Sakou et al.
patent: 5742504 (1998-04-01), Meyer et al.
patent: 5969753 (1999-10-01), Robinson
patent: 6122397 (2000-09-01), Lee et al.
patent: 6130967 (2000-10-01), Lee et al.
patent: 6141464 (2000-10-01), Handley
patent: 6148099 (2000-11-01), Lee et al.
IBM Technical Disclosure Bulletin, Jun. 1994, vol. 37, Issue 6B, pp. 399-402.*
Lee, JSJ, Haralick, RM, Shapiro, LG, “Morphologic Edge Detection”, IEEE Trans. Robotics and Automation RA3(2):142-156, 1987.
Serra, J, “Image Analysis and Mathematical Morphology”, London:Academic, 1982, pp 318-321.

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Structure-guided automatic learning for image feature... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Structure-guided automatic learning for image feature..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Structure-guided automatic learning for image feature... will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3024726

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.