Details:
Ex parte Bhatt
Appeal 2014-009532, Appl. No. 13/182,409, Tech. Center 2600
Decided: June 13, 2016
The application on appeal was directed to image processing, and more specifically, to presenting zoomed views of multiple faces. Each particular zoom level maximizes the fraction of the image occupied by the face. As explained in the application:
... it would be desirable to present zoomed views of multiple human faces depicted in the images 350-A, 350-B, 350-C and 350-D to allow a user associated with the system 300 to determine one or more faces from among the multiple faces are in focus or otherwise desirable and the image(s) from among the images 350-A, 350-B, 350-C and 350-D corresponding to the determined faces. To this effect, the GUI 302 can also include a control 320 through which the user can request that the system 300 concurrently zooms to portions of the multiple images 350-A, 350-B, 350-C and 350-D depicting faces of several persons.
... Upon receiving user input via the arrows of the control 320, the system 300 can switch, based on a display index/order associated with an anchor image, from concurrently presenting zoomed views associated with the detected instances of a person's face to concurrently presenting zoomed views associated with the detected instances of another person's face.A representative claim on appeal read:
1. A method performed by one or more processes executing on a computer system, the method comprising:The Examiner rejected the independent claims as obvious. According to the Examiner, the primary reference Shiritani taught most of the zoom limitations. In discussing the feature "zoom level selected to maximize a fraction ... occupied by the face," the Examiner relied on Shiritani's teaching "... in the case of an upright full-length figure image or the like, because the face area is small, the user has to observe the face area by enlarging the full-length figure image to be full-screen."
presenting, on a user interface viewable on a display coupled to the computer system, one or more digital images;
receiving user input requesting to zoom to faces depicted in the one or more digital images, where the faces include either human faces or animal faces; and
in response to receiving the user input and for each of the one or more digital images, automatically presenting on the user interface a zoomed view of the digital image
that displays an enlarged representation of a portion of the digital image showing an instance of a face depicted in the displayed portion of the digital image,
the zoomed view of the digital image covering an [entire] area of the user interface configured to present the one or more digital images,
the zoomed view being presented at a zoom-level that is selected to maximize a fraction of the displayed portion of the digital image occupied by the face.
The Examiner acknowledged that the primary reference did not teach zooming on multiple faces in response to user input, or doing so automatically. The Examiner turned to Nose for the zoom-in-response-to-user-input feature, the rationale for combining being "in order to allow users more flexibility in image zooming." Finally, for the feature of "automatically presenting ... a zoomed view," the Examiner relied on In re Venner, explaining that "it has been held that broadly providing an automatic or mechanical means to replace a manual activity which accomplished the same result is not sufficient to distinguish over the prior art." (MPEP 2144.04.III.)
The Applicant appealed and made multiple arguments against obviousness. With respect to the "maximize a fraction ... occupied by the face" limitation allegedly taught by the "full-screen" discussion in Shiratani, the Applicant argued that the reference did not teach making the face full-screen, but rather making the entire image full-screen. The Applicant then focused on the claim limitation "zoomed view covering an area of the UI configured to present the ... digital images," which the Examiner read on FIG. 9B of Shiratani.
The Applicant suggested that since Shiratani's UI included multiple images in the same window, the zoomed view did not cover the area presenting the digital image.
Finally, the Applicant traversed the Examiner's reliance on Nose for image zooming in response to user input. The Applicant pointed out that Nose described a two-step shutter button in a digital camera, where a half-press of the shutter button zoomed in on a face, and a full-press recorded the current image. Thus, zoom feature in Nose was a zoom before an image is captured, where the claim referred to zooming in a digital image already in existence.
In the Answer, the Examiner responded to the Applicant arguments with additional explanation. First, the Examiner reiterated that "Paragraphs 0017-0020 & Figures 9A/9B [of primary reference Shiratani] disclose maximizing or zooming an image which is already occupying an entire area of the user interface." The Examiner then explicitly explained how the combination was used to reject the claim: "Shiratani ... discloses enlarging the image after capture [while] Nose's contribution surrounds the ability to use the half-press [by the user] to zoom" (emphasis added.)
In a Reply Brief, the Applicant noted an inconsistency in the Examiner's position.
... [t]he Examiner has also made competing assertions with respect to Shiratani ... the Examiner has explicitly acknowledged that "[p]aragraphs 0017-0020 & Figures 9A/9B disclose an album with different faces being disclosed in two or more different images." ... In doing so, the Examiner has contradicted his previous assertion [in the Answer] that Shiratani's "[p]aragraphs 0017-000 & Figures 9A/9B discloses maximizes or zooming an image which is already occupying an entire area of the user interface.The Board agreed with the Applicant that the Examiner's reliance on Nose was in error.
(Emphasis in original.)
Nose's zooming by half-pressing a shutter button occurs before capture of an image with its "image taking device," whereas the claims are directed to post-capture processing of images with a computer system.Because the Examiner also admitted that Shiratani failed to teach this feature, the Board reversed the obviousness rejection.
My two cents: The Board got this one wrong: the Examiner's Answer clearly stated that it was the combination of Shiratani and Nose that taught the claimed feature. Nose provided a generic zoom-in-response-to-user-action, while Shiratani taught the claimed zoom details.
The Examiner made it easy for the Board to reach this conclusion, by presenting disjointed findings instead of a single coherent rejection. That is, the body of the rejection in the Answer explicitly stated that Nose taught the entire zoom limitation – and apparently that's the part the Board read. Then, the Response to Arguments section of the Answer contradicted the earlier statement, and explained instead that Nose taught zoom-in-response-to-user and Shiritani taught all other zoom details.
This habit – never revising a position, instead always piling on new stuff – is common practice by both Examiners and Applicants. I certainly appreciate it when the Examiner specifically responds to my arguments, but prefer that the Examiner do so by updating the rejection to take my arguments into account. If the Examiner instead reiterates the errored rejection then provides new information in a "Response to Arguments" section – as he did here – then the Applicant has two choices. One, file a response that goes point-by-point to address each contested Examiner statement as a separate argument. Alternatively, the Applicant could synthesize the total of all the Examiner's findings, resolving any internal contradictions, and make a coherent argument against this synthesized rejection.
Most Applicants take the first approach, even though it results in a disjointed argument – and IMHO a disjointed argument is by definition weaker than one that is coherently presented. I imagine there are a variety of reasons why Applicants don't incorporate multiple (and sometimes contradictory) Examiner findings into a single "synthesized" rejection. One, the Examiner contradicting himself is a good thing, right – highlight the contradiction and characterize it as error. Two, the Applicant's job is only to traverse the specific rejection on the record. I have a different viewpoint. As far as I know, an internal inconsistency or contradiction in a rejection isn't fatal error, so I don't see this as a strong argument. Moreover, assuming I'm confident about the "synthesized" rejection,* I am addressing the rejection on the record –I'm just doing a better job of explaining it than the Examiner did.
*I limit this "synthesized rejection" approach to instances where I am confident that I can resolve the internal contradictions/inconsistencies in the rejection. Ex parte Bhatt seems like a good candidate.