The shapes are defined by a binary mask outlining the objects.
MPEG7 CE Shape-1 Part B (database containing 1400 binary shape images)
closed binary shapes are collected from
- The Big Box of Art 615,000, Hemera Technologies
- The Big Book of Silhouettes, by Carol Belanger Grafton (Dover)
- articulated shapes (silhouettes)
hand-labeled masks Due to the labour involved in hand-labeling all image-to-image matches, the masks are approximate KTH Object Database (KOD)
- to provide an object database containing table top scenarios with more or less common household objects for attention and segmentation.
- to provide objects with differing complexity regarding shape and appearance, as well as scenes with different complexity regarding numbers of objects and object positions.
FRR_mask_<name>.pgm | Hand segmented mask of object <name> corresponding | to the right rectified foveal image Leaf shape image database a collection of leaf images from variety of plants (now 18 different types of leaf images)
• Content based image retrieval
• Non linear shape analysis
• Image segmentation for proper extraction of venation pattern
• Pattern classification
• Shape feature representation
• Application of curvature scale space
• Classification of images based on texture
• Venation pattern analysis
• Contour analysis
bullseye rating each image is used as reference and compared to all of the other images. The mean percentage of correct images in the top 40 matches (the 40 images with the lowest shape similarity values) is taken as bullseye rating. to generate a binary mask indicating the location of shapes in an image: img = imread('text.png'); radius = 1; imga = imdilate(img, strel('disk', radius)); %A bigger radius removes more details. imgb = imerode(img, strel('disk', radius)); This is the same as a morphological closing, imclose. You could also use a median filter to remove small objects and irregular boundaries. In a BW image, the median image would act like a majority filter.