1D Signals Descriptors for 3D Shape Recognition

In this paper, we propose a new 3D shape recognition approach. This approach is based on the shape recognition of a 3D object based on the processing of 1D signals, in order to reduce calculation complexity. The recognition is based on the calculation of shape descriptors from 1D signals. The first step in our approach is to convert 3D shape into 1D signals using a multi-line projection. These signals represent information on the third dimension of the object (Z). Then, the next step consists in calculating the 1D descriptors of these signals. These descriptors are used as input data of the classifier based on Euclidean distance to recognize the 3D object. The results of testing the proposed approach give an accuracy of 99.1%. This approach offers a simple, fast and efficient 3D shape recognition methodology, which makes our approach competitive for real-time applications.