データセットの違いが物体認識に与える影響の解析—特徴ベクトルの一致検索を用いた認識手法の場合—

要約

Specific object recognition methods based on exact matching of feature vectors are known as one of methods which can achieve high recognition performance for large-scale 3D specific object recognition. Since there are few common 3D object datasets whose size is sufficient to explore the effect of difference of object dataset composition and the effect of increasing number of objects for recognition, these effects for specific object recognition methods based on exact matching of feature vectors are discussed insufficiently. The number of objects in famous datasets (e.g., COIL-100) is around 100. Therefore, in this research, we prepare the dataset of 1002 3D objects by ourselves. In this paper, we will discuss the effect of dataset differences, which are based on object structure, texture and the number of objects, for those methods such as the method based on the Bloomier filter and the method based on a hash table with this dataset in addition to COIL-100.