Modelisation based on Rigdes and Blobs

We developed a new method for object representation based on ridges and peaks detected at several scales. Each object is represented as a graph such that each node is a feature (ridge or peak) and each arc is built from covering relation between spatial extensions of two features. This graph describes global shape as well as details of the object and allows many efficient strategies for graph matching.  

This figure indicates that ridges and blobs can be used to represent people structures. Ridges represent torso, aims, legs of a person.  Blobs represent round structure like head.

We apply our proposed modelisation of person based on ridges and blobs to classifiy people from non-people in sequence video (see site of CAVIAR for more detail) 

Our method is compared with two other methods for deteting people in the image, the one based on histogram of gradient and the another one based on auto-associative memory. These two methods are statistics, so do not give strutural representation of the object. We can see that our méthod is comparable. In addition, it can alse be applied to count the number of peoples in a group, that the two statistic methods can not.