The Test On 99shape Dataset - 924 Words

which classified into nine classes, with eleven shapes in each. Shapes in the same class are in different variant form, including occluded, noised, rotated, etc. Other databases including MPEG-7 Shape Dataset [5], Articulated Dataset, Swedish Leaf Dataset and Brown Dataset are used to have further experiments. Similar to [13], Precision and Recall is used for benchmark for the reason of fair comparisons. C. Results and Discussion Table I shows the optimal result from test on 99shape dataset. The numbers of points we sampled from the shapes are 50, 50 and 25 for RSD, RAD and TF respectively. For the articulated dataset, 45, 35 and 45 points are sample for RSD, RAD and TF. Retrieval result on articulated dataset was presented in Table I. We have noticed that result on 99shape from RAD is slightly better than RSD, while on articulated dataset RSD performs slightly better than RAD. During the above experiment, we tried to normalize the descriptors and found that experiment on 99 shapes received little influence from normalization while result from articulated dataset has some improvement. From the Table I, our algorithm has a almost 100% correct classified rate for Human and Wrench. We noticed that the Airplanes class is of the lowest correct rate except for the top 3 ranks. And the hit rate declined rapidly which make it singled out from the Table I. The matching distance in this class is carefully investigated and the distance revealed that our descriptorsShow MoreRelatedStandardized Databases And Benchmarks For Experiment979 Words   |  4 Pagesdifferent methods may have different assumptions on shapes. A commonly used database is 99shapes, by Kimia et al. It contains ninety nine planar shapes which classified into nine classes, with eleven shapes in each. Shapes in the same class are in different variant form, including occluded, noised, rotated, etc. Other databases including MPEG-7 Shape Dataset [5], Articulated Dataset, Swedish Leaf Dataset and Brown Dataset are used to have further experiments. Similar to [13], Precision and Recall is used