Lytro says its light-field camera, due to ship later this year, will let people change the focus of their photos after the fact.

A start-up called Lytro hopes to revolutionize photography by selling a camera later this year that lets people focus their images after the fact.

The technique used is called light-field photography, and it's been an active area of research for years in the optics realm. With it, lens and image sensor technology doesn't focus on a particular subject, but instead gathers light information from different directions' processing after the fact means different aspects of the scene can be recreated.

Lytro has been working on the technology for years--I interviewed Chief Executive Ren Ng three years ago when his start-up was called Refocus Imaging, and he began his research at Stanford well before that. But yesterday the company announced it plans to actually sell its first camera this year. Ng told All Things Digital's Ina Fried that the camera will be pocketable and "competitively priced," but was cagey on further details.

The promise of light-field photography is that people can fix or modify their photos afterward, for example focusing attention on a foreground subject by letting the background go blurry. Photographers have done this for years by setting a camera and lens for a particular depth of field and focusing, but Lytro argues its technology removes the technical challenges.

I foresee a number of challenges for Lytro itself, though.

First, disrupting the camera industry is hard. The digital camera revolution has increased the number of players from the old days of film, and there's already an overabundance of competitors. Companies such as GoPro show that it's possible to carve out a real niche, though, and even established players have fumbled with new challenges such as the network connections that make smartphone photography so powerful.

A second challenge is that near-field photography isn't a free lunch. It captures more depth at the expense of resolution. There's plenty of resolution to go around these days, one might argue. But Lytro will have to prove that its image quality is up to snuff.

Third, Lytro will have to train people to fiddle with photos after the fact, either in the camera or on a computer. E-mailing a JPEG or posting it to Facebook won't be as immediate, though. Lytro makes a virtue of necessity, calling its images "living photos" that let people explore the focus of the shot, but thus far it uses Flash Player and is far less convenient than a neatly packaged and universally readable JPEG. On my iPad, the "living photos" appeared to be static JPEGs.

There is something nice, though, about the idea of not having to worry about focus. People buy relatively expensive SLR cameras because they're frustrated by point-and-shoot cameras' long delay between pressing the shutter button and actually taking the photo. As the camera hunts for focus, the adorable smiling baby stops smiling or the slide at home plate is gone.

A parallel can be found among enthusiasts: raw photo formats. Higher-end cameras can merely record the image sensor data as a raw file, letting people process the photo after the fact with software. The approach offers greater flexibility and image quality, but involves more hassle. It seems to me that shooting raw is increasingly common, but I don't see any signs that the mainstream will embrace it anytime soon.

Lytro's approach also offers something powerful: it taps into the idea of computational photography, in which computers move beyond the optical limitations of cameras. Real-world examples include corrections for lens flaws including distortion, which can make parallel lines bow inward or outward, and vignetting, which can darken corners of an image.

Light-field photography is an extreme example of this trend. In effect it replaces complicated, expensive, tightly engineered optics of today's cameras with data processing. And at least in the last few decades, Moore's Law has shown more impressive progress than camera lens technology.

Lytro, based in Mountain View, Calif., brings the Silicon Valley entrepreneurial ethos to the idea. It's got 44 employees and has raised $50 million in three rounds, most from a third round from Andreessen Horowitz. Taking on the likes of Canon and Sony won't be easy, but the camera market is big enough that Lytro doesn't need to dethrone them to succeed.