Abstract:
The paper deals with matching several images of the same object and obtain the summary (average) of the image. On simple examples show that this average image itself cannot serve as a good benchmark for comparison with the recognition. It is proposed to use more of the summary image recognition. Experimentation with different methods of its construction and evaluation of their performance.
Keywords:image comparison function, the optimal position of the figures, Hamming distance, Manhattan distance, discrete Fourier transform.