Analyzing the results
This example presents a basic implementation that can be adapted in several cases such as 3D object recognition, face recognition, or image clustering. The goal of this chapter is to present how we can easily compare time series without any previous training, in order to find the similarity between images. In this section we present seven cases and will analyze the results.
In the following figure, we can see the first three searches and can observe a good accuracy in the result, even in case of the bus the result displays the result elements in different angles, rotation, and colors:
In the following figure, we see the fourth, fifth, and sixth search, and we can observe that the algorithm performs well with an image that has a good contrast in colors.
In case of the seventh search the result is poor, and in similar cases when the references time series is a landscape or a building, the result are images that are not related to the search criteria. This is because the...
