The object recognition is an important research area about computer vision. In the object recognition, the BoF (Bag of Features) is kind of generally technique and this is based on "bag of words". The BoF is a frequency histogram, is made from the local features of an image. This is not based on the feature distribution on the frequency histogram and this technique is using k-means and k-nearest neighbor. Therefore, our approach used the FCM (Fuzzy C-Means). The reason is a frequency histogram based on feature distribution for improving the recognition accuracy. The FCM uses another process about a frequency histogram than k-means and k-nearest neighbor. The FCM is a clustering method using fuzzy theory. This method allows variations of a local feature to make two or more clusters. The belonging level of each cluster is fuzzy membership. We discuss about our methods compared to the BoF's results. There results is made from two proposal method and one basic method. Finally, we compare about which category recognition. © Springer International Publishing Switzerland 2014.
CITATION STYLE
Shinomiya, Y., & Hoshino, Y. (2014). Bag of Features Based on Feature Distribution Using Fuzzy C-Means. In Communications in Computer and Information Science (Vol. 434 PART I, pp. 546–550). Springer Verlag. https://doi.org/10.1007/978-3-319-07857-1_96
Mendeley helps you to discover research relevant for your work.