Obviousness

To analyze the obviousness of US patent 6838651 under 35 U.S.C. § 103, we need to consider whether the differences between the claimed invention and the prior art would have been obvious at the time the invention was made to a person having ordinary skill in the art (PHOSITA). The filing date of US6838651 is March 28, 2002.

A PHOSITA in the field of CMOS image sensors at that time would likely have a strong understanding of semiconductor device physics, analog and digital circuit design, image processing fundamentals, and color imaging techniques, including common color filter array patterns like the Bayer pattern. They would also be familiar with the challenges and advantages of both CCD and CMOS imaging technologies, particularly concerning noise reduction, power consumption, and frame rates.

The framework for determining obviousness involves:

1. Determining the scope and content of the prior art.
2. Ascertaining the differences between the claimed invention and the prior art.
3. Resolving the level of ordinary skill in the pertinent art.
4. Evaluating any secondary considerations of non-obviousness (though the examiner's initial burden is to establish a prima facie case without these).

Prior Art References for US6838651:

The Google Patents page for US6838651 lists the following citations as "Prior art" (which typically means cited by the examiner or third parties):

• US3971065A (Eastman Kodak Company) - Color imaging array: This patent, filed in 1975 and issued in 1976, discloses a color imaging array with luminance-sensitive elements (green) dominating and chrominance-sensitive elements (red and blue) interlaid in a repeating pattern. It mentions using filters selectively transmissive to red, green, and blue light for producing luminance and chrominance-sensitive elements. It specifically discusses "color image sensors" for "video cameras" and different patterns for color filters.
• US5461425A (Stanford University) - CMOS image sensor with pixel level A/D conversion: This patent, filed in 1994 and issued in 1995, is generally relevant to CMOS image sensors with A/D conversion at the pixel level.
• US6380880B1 (Pixim, Incorporated) - Digital pixel sensor with integrated charge transfer amplifier: This patent, filed in 2001 and issued in 2002, describes a digital pixel sensor with an integrated charge transfer amplifier.
• US6611289B1 (Yanbin Yu) - Digital cameras using multiple sensors with multiple lenses: This patent, filed in 1999 and issued in 2003, discusses digital cameras that utilize multiple sensors and multiple lenses.

Analysis of Obviousness under 35 U.S.C. § 103:

Claim 1: A solid state imaging device, comprising: a red pixel having an output; a blue pixel having an output; a first green pixel having an output; a second green pixel having an output; a first analog-to-digital converter connected to the output of the red pixel for converting the output of the red pixel into a first digital signal and connected to the output of the blue pixel for converting the output of the blue pixel into a second digital signal; a second analog-to-digital converter connected to the output of the first green pixel for converting the output of the first green pixel into a third digital signal and connected to the output of the second green pixel for converting the output of the second green pixel into a fourth digital signal; and a color interpolation circuit for combining the first, second, third and fourth digital signals.

Motivation to Combine and Obviousness:

A PHOSITA at the time of the invention (2002) would understand the advantages of CMOS image sensors, including lower power consumption, higher levels of system integration (camera-on-a-chip), and the ability to support very high data rates due to parallel operation, as explicitly stated in US6838651 itself. They would also be aware that CMOS image sensors can easily be equipped with on-chip A/D converters.

The core of Claim 1 involves:

1. A pixel array with red, blue, and two green pixels (consistent with a Bayer pattern, which is explicitly mentioned in US6838651 as the basis for defining a color image).
2. Multiple A/D converters.
3. Specific allocation of pixel outputs to A/D converters (one A/D for red and blue, a second for the two greens).
4. A color interpolation circuit.

Combination of US3971065A and US5461425A to render Claim 1 obvious:

• US3971065A discloses a color imaging array using red, green, and blue filters, with luminance (green) elements dominating and chrominance (red and blue) elements interlaid in a repeating pattern. This directly teaches the pixel array structure of red, blue, first green, and second green pixels, as these are the components of a Bayer pattern for color image capture. The concept of individual pixels providing output signals proportional to incident light is fundamental to imaging devices.
• US5461425A teaches a CMOS image sensor with pixel-level A/D conversion. While it might not explicitly show two A/D converters specifically assigned to R/B and G/G channels, it establishes the concept of integrating A/D conversion within a CMOS image sensor architecture. A PHOSITA, recognizing the goal of achieving high frame rates and improved noise performance in CMOS sensors (as acknowledged in the background of US6838651), would be motivated to utilize multiple A/D converters to process the color information in parallel. The patent itself notes that "CMOS solid-state image sensors utilize multiple amplifiers that allow a longer settling time between applications and higher frame rate while maintaining excellent noise rejection" and "CMOS solid-state image sensors may easily be equipped with a precision analog-to-digital (“A/D”) converter on the solid-state image sensor chip". Furthermore, the patent describes that "two A/D converters may be employed, where one A/D converter is used for the red and blue channels and the second A/D converter is used for the green channels. In this manner, there is no addition to the fixed pattern noise of the imager that would arise from mismatch or offset in the two A/D converters."

A PHOSITA would find it an obvious design choice to group the red and blue pixel outputs to one A/D converter and the green pixel outputs to another, especially given the desire to process color channels independently for gain and offset adjustments (as mentioned in US6838651's summary) and to manage data rates. The patent explicitly states this as an alternative to using a separate A/D for each color channel.

The inclusion of a "color interpolation circuit for combining the first, second, third and fourth digital signals" would also be obvious to a PHOSITA. As stated in US6838651, "Color interpolation is used to determine the amount of red, green and blue light incident on each pixel. This process averages the color outputs of appropriate neighboring pixels to approximate each pixel's unknown color data." This is a standard image processing technique for demosaicing Bayer pattern images, well-known in the art at the time of the invention (US6330029B1, for example, discusses CFA patterns and the necessity to interpolate values).

Therefore, the combination of US3971065A (for the pixel array and color filtering) and US5461425A (for on-chip CMOS A/D conversion) would have made Claim 1 obvious. The motivation would be to create a higher performance CMOS image sensor capable of processing color information efficiently, with multiple A/D converters handling the data from different color channels, a recognized advantage of CMOS technology.

Claim 13: A solid state imaging device, comprising: groups of pixels, wherein each of said groups of pixels include: a red pixel having an output; a blue pixel having an output; a first green pixel having an output; and a second green pixel having an output; a first analog-to-digital converter connected to the output of the red pixel for converting the output of the red pixels into a first digital signal and connected to the output of the blue pixel for converting the output of the blue pixels into a second digital signal; a second analog-to-digital converter connected to the output of the first green pixel for converting the output of the first green pixels into a third digital signal and connected to the output of the second green pixel for converting the output of the second green pixels into a fourth digital signal; and a color interpolation circuit for combining the first, second, third and fourth digital signals.

This claim differs from Claim 1 primarily in emphasizing "groups of pixels" and the A/D converters being connected to the "outputs of the red pixels" (plural), "blue pixels" (plural), etc. This merely describes an array of the pixel groups defined in Claim 1. The obviousness arguments for Claim 1 apply directly to Claim 13, as the concept of an array of color pixels (as taught by US3971065A) combined with multiple A/D converters for processing different color channels (as motivated by the advantages of CMOS technology and the teachings of US5461425A) would be readily apparent to a PHOSITA. The generalization from individual pixels to "groups of pixels" in an array, where each group has the specified color components, is a straightforward scaling of the concept, especially when considering "solid-state image sensors" which are inherently arrays of light-detecting elements.

Claim 18: An imaging method comprising: converting an output of a red pixel into a first digital signal using a first analog-to-digital converter; converting an output of a blue pixel into a second digital signal using the first analog-to-digital converter; converting an output of a first green pixel into a third digital signal using a second analog-to-digital converter; converting an output of a second green pixel into a fourth digital signal using the second analog-to-digital converter; and combining the first, second, third and fourth digital signals using a color interpolation circuit.

This claim describes the method corresponding to the device of Claim 1. The obviousness of the method follows directly from the obviousness of the device. If the device described in Claim 1 is obvious based on the combination of prior art references, then the method of operating that device in the manner described would also be obvious to a PHOSITA. Converting analog pixel outputs to digital signals using A/D converters is a fundamental operation in digital imaging, and distributing these conversions across multiple A/D converters for different color channels is a logical design choice for improving performance, as previously discussed. The subsequent color interpolation is also a well-known process in digital color imaging.

Conclusion on Obviousness:

A prima facie case of obviousness can be established for the independent claims of US6838651 by combining the teachings of US3971065A and US5461425A.

• Motivation to Combine: A person of ordinary skill in the art would have been motivated to combine the color filter array technology of US3971065A (which defines the pixel arrangement for red, blue, and two greens) with the on-chip CMOS A/D conversion technology of US546145A. The motivation stems from the recognized advantages of CMOS imagers over CCDs in terms of power consumption, integration, and data rates, particularly the ability to employ multiple A/D converters for higher frame rates and improved noise performance, as explicitly stated in the background and summary of US6838651. The specific allocation of R/B to one A/D and G/G to another is a logical step for independent channel processing and managing data throughput in a CMOS sensor. The use of color interpolation is a well-established technique for reconstructing full-color images from such arrays.

The combination of these references addresses the problem of achieving high frame rates and improved sensitivity in CMOS imagers by efficiently converting and processing color pixel data.