AuthenTec ... another Apple acquisition


Image source: AuthenTec

Moving beyond the basis for crafting a title rife with alliteration there are interesting aspects to Apple's latest acquisition.  By and large Apple’s recent acquisitions seem to have been associated with technology destined for their A(x) processors. The most recent example being anobit and their flash controller technology.  Does AuthenTec fit this mold?  This deal seems a bit different because a large part of it involves technology around the actual fingerprint sensor. So, the answer is definitively...yes and no.  Let’s drill down a bit and see what's there.

It might be obvious, but a fingerprint sensor detects or images the pattern of ridges and valleys on your finger that combine to create a fingerprint.   Probably the most common approach measures variations in capacitance between areas of ridges or valleys.  In particular the sensor is organized as a grid of pixels, where each pixel contains a sensing capacitor, whose capacitance is a function of the presence of a ridge or a valley.  Another approach would be to optically image the finger, again detecting the ridges and valleys, but through light and shadow.

A brief look at a list of US patents assigned to AuthenTec finds the majority are related to the sensor.  At its most fundamental neither a capacitance nor optical approach is a new concept.  That said we should see patents related to improvements or refinements thereto.  As examples, US 8,115,497 discloses a sensor circuit for improving the data signal over the so-called “common mode” or background signal and US 8,180,120 discloses a system using polarized light. Beyond the actual sensor there should be systems and methods for processing the data.  Again taking two exemplary patents from the list, US 8,145,916 discloses systems and methods associated with encryption and US 7,734,074 discloses approaches for resampling image data as might be used in a so-called swipe sensor.  As with all patents these examples disclose particular aspects of technology, not actual products.

So, what about the products? There are plenty of discussions, here and here for example, out there related to the services that might be realized with the authentication provided by a fingerprint sensor.  But, how might Apple integrate the technology into their products? Apple is of course known for design, which is not friendly to a dedicated, "add-on" sensor.  It is also known for very particular performance criteria, evidenced by sometimes frustratingly withholding features until satisfied with the end user experience. 

In the iOS devices there just does not seem to be sufficient space for a dedicated sensor.  Mac’s may be a bit more forgiving in terms of space, but the addition of a sensor would be an unsymmetrical extra “thing”.  This would not likely be acceptable.  Rather, it is more likely that all efforts would be made to incorporate a sensor into the existing design.  This would be the touch display for the iOS devices and the trackpad for Mac devices.  It is not clear if these components currently have the resolution required to function as a fingerprint sensor, but a solution to this problem may come through evolution or maybe from within the AuthenTec IP.  One could envision “systems and methods for integrating a capacitance based sensor within a display” as an interesting patent title.

Any sensor requires some sort of processing.  Looking at some product pages on the AuthenTec website one finds sensors and sensor + chipset combo’s.  The chipset combo page in part reads, “This chipset maximizes first swipe success due to the rapid response and high performance possible by eliminating processing delays common when dependent on the host processor” teaching away from the use of existing or host processing capabilities.  Now it gets more interesting.

It is very unlikely that Apple would heed the above advice and provide a dedicated processor for processing a fingerprint.  It is far more likely they would go the host processor route.  For the Mac, there is likely enough horsepower in the current multi-core SandyBridge  processors, let alone what will be available at the time of any fingerprint integration.  What about iOS devices?  The last thing that Apple would accept here is a delay or, worse, a required second scan.  The ARM core of the A(x) processors has similar horsepower to the advertised sensor chipsets, but the ARM cores might also be loaded with other tasks at the time a fingerprint computation is desired.  However, there may be another option.  Maybe this task could be offloaded to one of the hardware engines that is otherwise designed for video.  It seems both tasks would require a highly repetitive calculation on each of many pixels.  It is further likely that such engines would be idle, in terms of video, at the time of a fingerprint capture.

 In the end it will be of interest to see how Apple moves forward with the acquired assets.  The key technology that prompted the takeover may already be public, or it could be lurking in a yet to be published patent application.  In the end, integration of this feature in the Apple design roadmap will be worth the wait.