While there's a giant crowd of people waiting for Diggto unveil its pictures section, there are already several services available for browsing popular news by pictures, including Google News and AOL's Mgnet (coverage). There's also a smaller, more independent group of sites that have been created by fans of some of these sites who have created their own visual solutions. One of them--Reddit Media--has increasingly become so popular it's led to Digpicz (Digg in pictures), and now PicUrls--which is a play on words, and similarity to the popular aggregator PopUrls.
The site pulls in stories and pictures from nine popular user-generated services including Digg, Flickr, Reddit, and Del.icio.us to name a few. It accomplishes this task by scraping the story data, going to the targeted news page, and pulling the relevant photo thumbnail. As the end user, you're getting a quick, direct link to the story, and a small view of what awaits you once you get there.
Like other aggregation services, PicUrls adds its own layer of social networking with integrated comments that layer on top of each service and integrated forums. You can also drill down to each individual site to see a complete listing of the latest popular stories, 16 at a time.
I absolutely love services like this that piggyback on one design and create a whole new way to visualize content. It's one of the things that made the Digg Labs visualization contest so enjoyable. While I'd prefer to see these projects using open APIs over scraping, in this case it's only being done once for the service, rather than for every user.