Obviousness

Obviousness Analysis under 35 U.S.C. § 103 for US8471950

This analysis identifies combinations of prior art references that would render the claims of US patent 8471950 obvious to a person having ordinary skill in the art (PHOSITA) as of the priority date (2008-11-28). The core inventive concept of US8471950, as exemplified by its independent claim, is to improve automatic image quality control by detecting features (e.g., spotlights), removing pixels associated with these features from an evaluated value extraction region, extracting evaluated values from the resulting modified region, and then using these "clean" values to control image processing.

Overview of the Invention (US8471950)

US8471950 describes a signal processor that includes:

1. A feature detection unit to identify positions where pixel features (e.g., sudden signal intensity changes indicating a spotlight) have changed.
2. A region creation unit that modifies a predefined evaluated value extraction region by removing the pixels with the detected features.
3. An evaluated value extraction unit that extracts values from this modified region.
4. A processing unit that applies image quality adjustments (e.g., exposure, focus, white balance).
5. A control unit that controls the processing unit based on the evaluated values obtained from the feature-removed region.

The patent explicitly highlights a problem with prior art, specifically JP-A-6-14236, stating that "when there are a plurality of high luminance portions, there is a problem that it is necessary to repeat the change of the region and the process of confirming the absence of the high luminance portion." This indicates a motivation to improve the efficiency and accuracy of evaluated value extraction by directly addressing the problematic regions.

Prior Art Combination and Obviousness Rationale

Primary Combination: JP-A-6-14236 in view of US20030002732A1 (Gossett) and/or general knowledge of image processing techniques.

JP-A-6-14236

As described in the background of US8471950, JP-A-6-14236 discloses:

• An object "to obtain safe and accurate focusing information from an image region (ranging frame) with exposure most suitable for AF and without high luminance."
• A solution where "a system control circuit changes an image region for detection of a focusing lens position when a high luminance detection circuit detects a high luminance portion."

From this description, JP-A-6-14236 teaches most elements of Claim 1:

• Feature detection unit: The "high luminance detection circuit" detects high luminance portions (features).
• Evaluated value extraction unit: The system obtains "focusing information" (an evaluated value) from an image region.
• Processing unit and Control unit: The system is for "AF" (automatic control) and involves a "system control circuit" changing regions for "detection of a focusing lens position," implying control over a focusing lens (processing unit) based on evaluated values.

However, US8471950 identifies a critical drawback in JP-A-6-14236: "the change of the region is not performed after the absence of the high luminance portion in the changed region is confirmed... Therefore, when there are a plurality of high luminance portions, there is a problem that it is necessary to repeat the change of the region and the process of confirming the absence of the high luminance portion." This problem demonstrates a clear motivation for a PHOSITA to find a more efficient and precise method of defining the evaluated value extraction region.

US20030002732A1 (Gossett)

US20030002732A1 (Gossett) is titled "Method and apparatus for digital image segmentation using an iterative method." Image segmentation is a well-known image processing technique for partitioning a digital image into multiple segments or sets of pixels. This technique is used to precisely identify and isolate specific portions of an image based on their features.

Motivation to Combine and Obviousness

A PHOSITA, aware of the image processing system in JP-A-6-14236 and its inefficiency when dealing with multiple problematic features (e.g., high luminance portions) within the evaluation region, would be motivated to improve this system. The goal would be to reliably exclude these problematic features from the evaluated value extraction region in a single, non-iterative step.

It would be obvious for a PHOSITA to combine the teachings of JP-A-6-14236 with known image processing techniques, such as image segmentation or masking, for the following reasons:

1. Known Problem and Predictable Solution: JP-A-6-14236 clearly articulates the problem of iterative region changes. Image segmentation, as taught by Gossett, provides a direct and well-understood method for precisely identifying and isolating specific pixels or regions based on features. Applying such a segmentation technique to the high luminance portions detected by JP-A-6-14236 would predictably allow for their direct removal or exclusion from the evaluation region, thereby addressing the stated inefficiency and improving the accuracy of the extracted evaluated values.
2. Standard Design Choice: Once a problematic feature is detected (as by JP-A-6-14236), the most direct way to prevent its influence on evaluation is to mask it out or segment it away from the region of interest. This is a fundamental concept in image processing for obtaining accurate measurements from specific parts of an image while ignoring others.
3. Efficiency Improvement: By directly removing the pixels with the detected features from the evaluation region, rather than iteratively changing the entire region, the combined system eliminates the need for repeated checks, thus increasing processing efficiency, which is a desirable goal for any automatic control system.

Therefore, a PHOSITA would be motivated to combine JP-A-6-14236 with the techniques of US20030002732A1 (Gossett) (or general image segmentation/masking knowledge) to implement a "region creation unit which creates a region by removing pixels with the features from an evaluated value extraction region." This renders Claim 1 obvious.

Obviousness of Dependent Claims

The dependent claims build upon the core invention by adding standard components, applying the invention to known applications, or detailing specific implementations.

• Claim 2 (Hold Unit): A "hold unit which holds position information of the detected pixels" is a standard functional component in any digital system where detected information (like pixel positions of features) needs to be stored for subsequent processing (e.g., by the region creation unit). Adding a hold unit to store the feature positions detected by JP-A-6-14236 and/or segmented by Gossett would be an obvious design choice for a PHOSITA.
• Claims 3 & 12 (Focus Lens Control): JP-A-6-14236 explicitly concerns "focusing information" and "AF," using the detected high luminance portions to modify regions for "detection of a focusing lens position." Applying the improved signal processor (JP-A-6-14236 modified with pixel removal) to the existing autofocus application described in JP-A-6-14236 would be an obvious continuation of its original purpose, aiming for more accurate AF. US20010028402A1 (Sanyo Electric Co., Ltd.) and US20060087578A1 (Samsung Techwin Co., Ltd.) further confirm that autofocus was a well-known field encountering such issues.
• Claims 4 & 14 (White Balance Adjustment): Adjusting white balance in digital cameras is a well-known function. Prior art such as US20040169767A1 (Norita) for a "Digital camera and control method thereof" or JP2008176211A (Hitachi Ltd) broadly address digital camera control which includes white balance. Strong light sources (like spotlights) can skew white balance calculations. Extending the feature-removal technique to white balance evaluation to obtain more accurate color balance would be an obvious application for a PHOSITA, analogous to its application in AF or exposure control, to prevent inaccurate readings from problematic image features.
• Claims 7 & 10 (Light Intensity Adjustment/Exposure Control): The problem of spotlights affecting evaluated values, which can lead to underexposure of other parts of the image, is a known issue in automatic exposure control. JP-A-6-14236 mentions "exposure most suitable for AF." Prior art like JP4077217B2 (Ricoh) specifically addresses "Automatic exposure control device." Applying the feature-removal technique to exposure control to prevent bright spots from distorting overall exposure calculations is an obvious extension for a PHOSITA to improve image quality.
• Claims 5 & 8 (Region Segmentation Unit and Overlap Detection Unit): These claims describe a specific implementation of the region creation unit using a "region segmentation unit for segmenting an effective region in advance" and an "overlap detection unit which detects overlap of position information output from the hold unit with blocks segmented by the region segmentation unit." US20030002732A1 (Gossett) teaches digital image segmentation, which can be implemented using various methods, including block-based approaches. The concept of segmenting an image into blocks and then identifying and excluding blocks that overlap with detected features is a common and obvious way to implement pixel-level removal for region creation in image processing.
• Claims 6, 9, 11, 13, & 15 (Keeping Total Area Constant): These claims involve keeping "the total area of the evaluated value extraction region constant by excluding unnecessary portions from the evaluated value extraction region." In automatic control systems, maintaining a consistent evaluation area can simplify control algorithms and lead to more stable and predictable adjustments, as the absolute magnitude of an evaluated value might otherwise change simply due to a change in the evaluation area. For a PHOSITA, compensating for removed areas by expanding other valid regions or adjusting the evaluation metrics to maintain consistency would be an obvious design consideration for improving control stability and simplifying subsequent processing.

In conclusion, the core innovation of removing specific pixels with detected features from an evaluation region, rather than iteratively changing the entire region, would have been obvious to a PHOSITA by combining the problem identified in JP-A-6-14236 with established image segmentation or masking techniques known in the art (e.g., from US20030002732A1). The applications to specific image quality adjustments (AF, exposure, white balance) and the detailed implementations (hold unit, segmentation by blocks, constant area) are also rendered obvious as logical and predictable refinements or applications for a PHOSITA.