Currently computers have difficultly with recognizing objects in images. While practical solutions exist for a few simple classes, such as human faces or cars, the more general problem of recognizing all the different classes of objects in the world (e.g. guitars, bottles, telephones) remains unsolved. Computer vision researchers are currently investigating methods that can recognize and localize thousands of different object categories in complex scenes. A key component of these algorithms is the data used to train the computer's model of each object. The goal of LabelMe is to provide an online annotation tool to build a large database of annotated images by collecting contributions from many people. You can contribute to the database by visiting the annotation tool (check our annotation guidelines).

Here, we display all of the annotated images in LabelMe, arranged as a montage of tiny thumbnails. The images are arranged according to semantic similarity between images (nearby images will contain similar objects in similar spatial configurations). When you click on one of the tiny thumbnails in the montage, a larger thumbnail will appear, along with a list of the labeled objects and a segmentation map showing their position in the image. You can also switch the display of the mosaic to view the images (click on the "Images" icon at the top right) or the annotations (click on the "Labels" icon).