In 2013, Tensilica (subsequently acquired by Cadence) released its second-generation image processing IP core, the IVP, which also supported modest computer vision capabilities (Figure 1).

Since reading Malcolm Gladwell's Blink a decade ago, I've been intrigued by how the mind works – particularly how judgements and decisions are made. I've been inspired to take an armchair tour of research on this topic, and have encountered fascinating insights from the likes of David Eagleman and Daniel Kahneman.

Reading the work of these talented researchers and writers has led me to the inescapable conclusion that most of our judgements and decision-making take place in our subconscious minds.

The decreasing cost-per-transistor delivered by modern semiconductor processes means that a number of previously rare embedded processor options are now increasingly common. This trend includes floating-point coprocessors, which are especially useful when migrating code originally developed on a PC to an embedded system.

In September, Freescale announced its acquisition of computer vision processor IP supplier CogniVue. BDTI discussed the news with Matt Johnson, Vice President and General Manager of Freescale's Automotive MCU division, which instigated the transaction. The two companies had closely collaborated for the past several years, so the purchase wasn't a complete surprise. Still, the interview produced a number of interesting insights.

In 2002, in a famous piece of unintentional rhetorical artistry, U.S. Defense Secretary Donald Rumsfeld spoke to reporters about "known knowns" (things you're aware that you know), "known unknowns" (things you're aware that you don't know) and "unknown unknowns" (things you're unaware that you don't know). As every programmer immediately realized, the category that Rumsfeld didn't mention is the "unknown knowns" – things you know but don't realize that you know.

A growing number of products are incorporating computer vision capabilities. This, in turn, has led to rapid growth in the number of processors being offered for vision applications.  Selecting the best processor (whether a chip for use in a system design, or an IP core for use in an SoC) is challenging, for several reasons.

It's becoming increasingly possible, thanks to APIs and languages such as Khronos' OpenCL, for applications to efficiently harness the heterogenous computing resources available in modern SoCs.

As Jeff Bier has mentioned in several of his recent columns, deep learning algorithms have gained prominence in computer vision and other fields where there's a need to extract insights from ambiguous data.

My colleagues and I at BDTI frequently evaluate processors – sometimes on behalf of processor suppliers, and sometimes on behalf of processors users. Comparing processors can be complicated, given their many attributes, including some (like energy efficiency) that are difficult to compare accurately, and others (like ease-of-use) that are inherently subjective.

Several recent conversations with processor suppliers have reminded me of another serious hazard in processor comparisons, one that I call the "Uncertainty Principle of Processor Comparison."