Thursday, June 9, 2011

Leafsnap



Peter Belhumeur is a professor of computer science at Columbia University and director of the Laboratory for the Study of Visual Appearance. It was his 8-year-old son William who suggested he create an application that identifies plants using the face recognition technology he has worked on since the mid-1990s. With the help of colleagues from Columbia, the University of Maryland, and the Smithsonian Institution, he has done just that! Using the same algorithms that can process the curves and angles of a human face, he designed a program that will narrow down (smooth versus jagged, many-lobed or single-lobed) and determine plant species based on the shapes of their leaves. LeafSnap, an electronic field guide very useful to botanists but easy enough for a child to use, is now available free* for the iPhone and iPad, and will be available for Android phones later this year. Users - and there are already 150,000 - photograph a leaf, upload it, and see a list of possible matches within seconds, along with other characteristics (flowers, fruit, seeds and bark) to help confirm the tree's identity. Below is the data that accompanies images of the plant I chose to feature:

Paper Mulberry

Broussonettia papyrifera

Native to

eastern Asia, this species is now widely naturalized in the eastern United States where it was introduced for its ornamental value. Its leaves are highly variable in shape, even on the same branch, but are usually 7-20 cm long with a rough surface above and fuzzy below. The juicy red fruits are edible and sweet, but are very fragile and not suitable for commercialization.
Habitat:Grows in a wide variety of habitats.

Growth Habit: Fast-growing tree to 12 m, with round crown, sprouts into thickets.

Bloom Time: Late spring.

Presence in US: AL AR CT DC DE FL GA HI IL IN KS KY LA MD MO MS NC NJ NY OH OK PA SC TN TX VA
Belhumeur’s student volunteers photographed 1,000s of leaves from the 145 species in New York's Central Park and the 160 species in Washington, D.C.’s Rock Creek Park. The database covers most native species in the northeast, with plans to add the additional species native to North America, and possibly other continents. “It was fun to take that visual recognition technology and drop it into this domain, working with biologists and doing something I cared about as a kid,” says Belhumeur. Speaking of which, his son is already dreaming up new apps. “I think it is time for Fishsnap and Bugsnap,” says William, “so there is still a lot of work to do.”

*Leafsnap cost about $2.5 million to develop, funded primarily by a grant from the National Science Foundation.

1 comment:

  1. A friend has the app that identifies songs based on input to the microphone. The face recognition app in Facebook is surprisingly good. Leafsnap is a terrific idea. The first thing that comes to mind is poison oak! It makes me wonder what is next in a world where we have such incredible access to information.

    ReplyDelete

You may add your comments here.