What are the classifications of graphics?

There are also types of plane, solid plane and 3d plane: circle, rectangle, square, parallelogram, triangle and trapezoid; Cylinders, cones, spheres, cuboids and cubes.

In order to make use of the characteristics of the intermediate band, Chang Heguo developed a tree structure wavelet transform to further improve the classification accuracy. Some researchers combine wavelet transform with other transforms to obtain better performance. For example, Thygaarajna and others combine wavelet transform with * * * generating matrix to take into account the advantages of statistical and transform-based texture analysis algorithms.

Extended data:

In fact, the more commonly used method is to classify shapes by combining regional features and boundary features. For example, Eakins and others put forward a set of redrawing rules and simplified the shape outline with line segments and arcs, and then defined two classification functions of adjacency family and shape family to classify shapes.

Adjacent families mainly use the boundary information of shapes, while shape families mainly use the information of shape regions. When matching shapes, besides the differences of shapes in each family, we also compare the differences of centroid and perimeter in each family and the differences of position feature vectors of the whole shape, and the query discriminant distance is the weighted sum of these differences.

Baidu Encyclopedia-Image Classification