WORD: Wiry Object Recognition Database

Image Sets:

Chair

Red Chair

Cart

Example image

Example image with labeled edges

Number Of Images: 174

Image Size: 1024 x 768

Comments: This image set shows a 4-wheeled push cart at arbitrary orientations with respect to the camera. The images span the full revolution of the cart in the plane parallel to the floor. Across all images, the elevation of the camera varies by approximately 25 degrees with respect to the object, and the extent of scale variation across images is about 10%. The height of the camera varies between 1.5 m and 2 m and the distance from the camera to the object is around 2-3 m. The background was segmented using the fill tool in Adobe Photoshop. The cart is in the lower left in all images to facilitate compositing with the clutter A, clutter B, and clutter C images.

Most edges detected by vista are from occluding contours, however some of the edges are generated by patterns of rust on the object (see above example image). In some images, the cart wheels are in deep shadow and are poorly segmented.

Vista Parameters: (sigma, lowThreshold, highThreshold,minimumEdgeLength, accuracy, granularity, magnitude) = (3,10,35,50,2.0,4.0,20.0)