'

Searching in 3-D

Eighteen months ago, I was writing that shape searching could become a reality. Now, the researchers at Purdue University who developed this initial system are providing benchmarking tools to evaluate how well their search system is working.

Eighteen months ago, I was writing that shape searching could become a reality. Now, the researchers at Purdue University who developed this initial system are providing benchmarking tools to evaluate how well their search system is working. Even if these tools are designed for computer-aided design (CAD) engineers, you're welcome to use the Purdue benchmarking system for free. Its database contains about 1,000 parts grouped in 40 categories. You select a part, such as tha backdoor of a car, and you can see how well the system can find other 3-D objects with similar shapes in the database. The idea of searching parts in a manufacturer database was born from the frustration of CAD users who routinely spend up to 80 percent of their time searching for information stored in the computers of their companies and trying to reuse it.

Here is the introduction of this Purdue University news release.

Shape-search engines could save time and millions of dollars annually by making it easier for companies to "reuse" previous designs, reducing redundancy and streamlining a company's supply chain. The systems will enable companies to benefit from the lessons learned in creating past parts, said Karthik Ramani, a professor of mechanical engineering and director of the Purdue Research and Education Center for Information Systems in Engineering.

So how does this search system work? First, you select a part in the inventory that looks similar to the one you need -- or want to design. Or can even use draw a desired part on your computer screen. Then the new benchmarking tools will help you to find if a previous object with a similar shape already exists, saving you time.

The Purdue benchmarking system uses an inventory of 1,000 parts and evaluates how well a search system is able to retrieve matches to a part entered into a query. The parts are grouped in 40 categories, such as ringlike parts, T-shaped parts, cylindrical parts and disk-shaped parts.

"If I give a query for a part that's in one of the categories, the top 10 results should ideally be in that category and as close to the queried part as possible," Ramani said. "If the search system found only six matches from the right category and four from some other category, then I know it's not that good."

Now, it's time to see how well these tools are working. First, let's go to the Shapelab. Now, let's select "a solid of revolution," then "bolt like parts," then a simple bolt.

Below is a partial screenshot of similar objects to this simple bolt found by the system (Credit: Purdue University).

Searching for a bolt

If this experiment doesn't convince you, feel free to use this tool for other trials. But remember that the database is very small compared to the huge ones maintained by corporations such as Boeing or DaimlerChrysler to name a few.

And even if these new tools are not perfect, they might be able to help to save millions of dollars annually for such companies.

The latest research works have been presented at the ASME 2005 International Design Engineering Technical Conference, but the papers have not been publicly released yet.

Sources: Purdue University news release, September 26, 2005; and various web sites

You'll find related stories by following the links below.