Performance Evaluation of Different Soft Clustering Algorithms for Bag of Words Model
Object categorization by Bag-of-Words (BoW), which represents an image of an object by the histogram of local patches on the basis of a visual vocabulary, has attracted intensive attention due to its good performance and flexibility. An object recognition method based on the Bag-of-Words model is implemented were descriptors are quantized to form a visual word dictionary called codebook with the help of different soft clustering algorithms. The performances of four different soft clustering algorithms are evaluated. These different algorithms are evaluated in terms of macro precision, micro precision, accuracy and F1 measure. The proposed modified algorithms give an optimal fuzzy partition by minimization of a modified objective function.