分析技术：让世界变得更安全、更清洁、更美好

方进（Gene Frantz）
德州仪器（TI）首席科学家

人们首次听说“视频分析技术”这个词时，许多熟悉这种技术的人都先联想到了视频安全应用。不过，分析技术的影响远远超过安防领域，它可以让我们享受到更安全、更清洁的生活，促进人与社会之间更具个性化的全新互动方式。正像我以前说的那样，存在无限可能。

我们先来谈谈分析技术对节能的作用。工程师们认识到，分析技术可提高能耗效率，降低碳排放量。例如，具备分析功能的智能摄像头可检测到房间中的人数及其所处的位置，发送该信息给空调系统，从而自动调节房间的气流与温度。同样，照明设置也可根据室内人员进行调节。就生物检测分析技术而言，空调与照明系统还能根据房间中人员的节能要求作进一步的微调设置。以上只是分析技术减少楼宇碳排放量的几个最简单的例子，这样的例子还有很多。

节能这个理念很好，但我们大多数人都不想为了节能而被摄像头全天候监视。这是一个可以理解的顾虑。我也不喜欢待在摄像头前，不信可以问问最近刚给我拍过照的摄影师。此外，人们不愿为大量存储视频节目而发愁。令人欣慰的是，大多数分析应用不需要存储视频内容，因为智能摄像头能处理输入信息，而不只是当作数据存储。下面的例子说明了数据与信息的差别。所谓信息，是指摄像头能了解房间中是否有人，而数据则是单纯拍下房间中有人的画面。就上述节能例子而言，摄像头能观测某个场景，作出决定，然后删除数据。我认为，也没什么人有兴趣去看摄像头拍下的内容，上述技术还使得我们不用手工去删除拍摄的节目。......

分析技术的未来

方进（Gene Frantz）
德州仪器（TI）首席科学家

可编程的智能摄像头提供高级智能分析功能，是支持新一代安全与汽车应用的重要技术。
智能摄像头将不断收集、处理我们周围的信息，并自动采取行动，这样我们就可抽出时间来做其它的工作。如果有人在凌晨一点翻过围墙闯入，那么摄像头将自动识别出这个人的身份，到底是自家的孩子偷偷回家，还是想把电视机搬走的小偷。在机场或其他公共场所，摄像头能判断出被人遗留的包裹，并联系相关管理部门。在街头，摄像头能减少警方与其它应急机构的工作量。在汽车中，摄像头也可发挥安全功能，甚至最终实现自动驾驶。
目前，我们考虑视频分析时，通常将最终用户设想为某个人。我们认为某个人会接收到一定存储容量的信息，其中包括各种压缩与复制数据。不过未来分析技术的最终用户将是一台电脑，到时候数据是否压缩将根本不重要了，图像是否清楚也无关紧要，对一台设备来说，某个像素不清楚就像一粒无足轻重的尘埃一样可以忽略。到时候，分析技术的重点不再是数据，而是以智能为中心，这是技术的一大进步，不仅涉及到技术过渡，还涉及到社会行为的变化。人们可能还不太习惯让机器在无人干预情况下作出重要决策。
我认为，上述技术过渡会分步骤进行。以汽车工业为例来说，首先，我们通过摄像头帮助司机做出更好的驾驶决策。比方说，摄像头能看到司机盲区的情况，告诉司机盲区内是否有车，司机就能根据这些信息作出决定。在技术发展的下一步，摄像头可以为决策提供参考意见。要是在发现邻车道上有车的情况下司机还要并线，那么就会自动加大方向盘的阻力，提醒司机潜在的问题与危险。第三步，摄像头将成为主要的决策者，同时给司机提供拒绝摄像头操作的选择。最终，我们将实现摄像头作出唯一决策的目标。
当然，实现所有上述发展都需要可编程性。原因何在呢？不仅因为摄像头必需作出无数决策，才能对系统发出指令，而且同时商业发展现实也要求可编程性。我们要求可编程摄像头能根据应用需求采用适当软件，并随时间的推移不断发展。不过根本因素是只有确保足够的可编程性，我们才能随心所欲地实现各种功能。
一旦实现了上述目标，智能摄像头就能应用于日常生活的各个领域。交通部门不仅可用它来检查哪里发生堵车，还可分析堵车的症结所在。摄像头能用于各种安全领域，包括机场的面部识别，保护医院里的新生儿等。
当然，未来人们的观念也需要改变。我们须信赖摄像头及其所作的决定。......

DSP’s Past Can’t Hold A Candle to its FutureBy Gene Frantz
TI Principal Fellow and Business Development Manager, DSP

Has it really been 25 years? Can it be possible that TI’s digital signal processing technology is old enough to rent a car and appear in a light beer commercial? In 1982, the semiconductor market saw the first viable digital signal processor shipped. Over the ensuing 25 years a paradox has arisen. As they’ve aged, DSPs have grown smarter, faster and more agile while I have grown slower, and more apt to spend a Saturday night in front of the TV. Tempus fugit.......

Forward an speech from Mike Hames

The Next 25 Years
By Mike HamesTI Senior Vice President
I had the chance this morning, at TI’s fifth annual Developer Conference, to get together with about 1000 engineers who are using TI Digital Signal Processing technology to create unique and exciting electronics products. As you may know, this is the 25th anniversary of the first commercially successful digital signal processing. And as someone who was there at the beginning, I am pleased to join the conference today to talk about where the technology is headed in the next 25 years.......

What is new with SoCs
By Gene Frantz
TI Principal Fellow and Business Development Manager, DSP

Another comment we received recently was from Shreshtha Kumar, questioning all of our hype around DaVinci technology and what is new about it. Again this is something, I wanted to talk to and as always I welcome your feedback on my thoughts. Kumar wrote, “SOC concept is not new. It is here for quite a few years but it complex nature sticts (sic) its rapid development. Here comes DaVinci for rescue.”......

What Moore Didn’t Tell Us about ICs
By Gene Frantz
TI Principal Fellow and Business Development Manager, DSP

Now that we’ve painted some broad strokes about SoCs and the applications that use them, I’d like to take a step back and fill in some of the gaps. Three questions that keep us up at night: What are the physics that are driving the move to SoCs? How do we stay ahead of the limitations of those physics? And, just what are SoCs, anyway? My next few postings will attempt to answer these.......

Is There a Consumer Breaking Point In Terms Of Convergence?

Question: Is technology driving the move towards convergence or are manufacturers and technology suppliers simply responding to artificial market demand and expectations? Answer: We can’t be positive one way or the other. All we know for sure is that consumers want better cameras, better video, VoIP, and WLAN, and they are running out of pockets. So it only makes sense to build convergent multifunctional devices. But there are also conflicting signals in the convergence mix. ......

How Small is Too Small?By Gene Frantz
TI Principal Fellow and Business Development Manager, DSP

As I have stated, the use of SoC will increase as we are able to pack more and more converging technologies into a single portable device. But simply because we can integrate doesn’t mean we should do it. It will not always be practical to do, nor will it be the right thing to do.

Beyond the fact that color displays are the de facto standard for all portable equipment, nothing else concerning displays for portable devices is certain. That leaves us with an important, fundamental question that must be answered: How small is too small? ......

SoCs Are Answering Demand for Converged Portable Devices
By Gene Frantz

TI Principal Fellow and Business Development Manager, DSP

In my last blog, I wrote about the trend for application designers to create new products on existing SoC platforms like TI’s DaVinci technology. While this seems to turn the supply/demand paradigm on its head, the opposite is true. Let me give you a couple examples.

The demand for multifunctional portable equipment featuring convergent technologies is driving several markets. Cellular telephones already incorporate personal-digital-assistant (PDA) features, global-positioning-system (GPS) functions, and take still photos and video clips. Soon, still-photo quality will improve exponentially and there will be an even bigger leap in video quality. Convergence will shift into high gear as the new models incorporate TV features, Wireless LAN (WLAN) and Voice over Internet Protocol (VoIP). ......

Architecture or Application?

By Gene Frantz
TI Principal Fellow and Business Development Manager, DSP

Are applications driving architecture design, or are architecture designs driving applications? Seems like a heretical chicken-and-egg question from a business standpoint. Our customers design applications, therefore their needs are paramount and our architectures should address those needs – they demand we supply: business 101.

But the realities of the semiconductor industry, namely the cost of creating new process nodes and fabrication – have made the answer to this question a little murkier. Today, for the most part, it is prohibitive for companies to design a chip geared toward a single application. This is especially true for smaller companies where, traditionally, innovation occurs. Therefore, innovation itself is in jeopardy if we adhere to old business models.......