Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval
Image retrieval based on a convolutional neural network (CNN) has attracted great attention among researchers because of the high performance. The pooling method has become a research hotpot in the task of image retrieval in recent years. In this paper, we propose the feature distribution entropy (F...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-10-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/21/11/1037 |
| _version_ | 1798038540514230272 |
|---|---|
| author | Pingping Liu Guixia Gou Huili Guo Danyang Zhang Hongwei Zhao Qiuzhan Zhou |
| author_facet | Pingping Liu Guixia Gou Huili Guo Danyang Zhang Hongwei Zhao Qiuzhan Zhou |
| author_sort | Pingping Liu |
| collection | DOAJ |
| description | Image retrieval based on a convolutional neural network (CNN) has attracted great attention among researchers because of the high performance. The pooling method has become a research hotpot in the task of image retrieval in recent years. In this paper, we propose the feature distribution entropy (FDE) to measure the difference of regional distribution information in the feature maps from CNNs. We propose a novel pooling method, which fuses our proposed FDE with region maximum activations of convolutions (R-MAC) features to improve the performance of image retrieval, as it takes the advantage of regional distribution information in the feature maps. Compared with the descriptors computed by R-MAC pooling, our proposed method considers not only the most significant feature values of each region in feature map, but also the distribution difference in different regions. We utilize the histogram of feature values to calculate regional distribution entropy and concatenate the regional distribution entropy into FDE, which is further normalized and fused with R-MAC feature vectors by weighted summation to generate the final feature descriptors. We have conducted experiments on public datasets and the results demonstrate that our proposed method could produce better retrieval performances than existing state-of-the-art algorithms. Further, higher performance could be achieved by performing these post-processing on the improved feature descriptors. |
| first_indexed | 2024-04-11T21:41:35Z |
| format | Article |
| id | doaj.art-e9c246abde45450a926bcd0a79ff2e8b |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-04-11T21:41:35Z |
| publishDate | 2019-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-e9c246abde45450a926bcd0a79ff2e8b2022-12-22T04:01:35ZengMDPI AGEntropy1099-43002019-10-012111103710.3390/e21111037e21111037Fusing Feature Distribution Entropy with R-MAC Features in Image RetrievalPingping Liu0Guixia Gou1Huili Guo2Danyang Zhang3Hongwei Zhao4Qiuzhan Zhou5College of Computer Science and Technology, Jilin University, Changchun 130012, ChinaCollege of Computer Science and Technology, Jilin University, Changchun 130012, ChinaCollege of Computer Science and Technology, Jilin University, Changchun 130012, ChinaCollege of Computer Science and Technology, Jilin University, Changchun 130012, ChinaCollege of Computer Science and Technology, Jilin University, Changchun 130012, ChinaCollege of Communication Engineering, Jilin University, Changchun 130012, ChinaImage retrieval based on a convolutional neural network (CNN) has attracted great attention among researchers because of the high performance. The pooling method has become a research hotpot in the task of image retrieval in recent years. In this paper, we propose the feature distribution entropy (FDE) to measure the difference of regional distribution information in the feature maps from CNNs. We propose a novel pooling method, which fuses our proposed FDE with region maximum activations of convolutions (R-MAC) features to improve the performance of image retrieval, as it takes the advantage of regional distribution information in the feature maps. Compared with the descriptors computed by R-MAC pooling, our proposed method considers not only the most significant feature values of each region in feature map, but also the distribution difference in different regions. We utilize the histogram of feature values to calculate regional distribution entropy and concatenate the regional distribution entropy into FDE, which is further normalized and fused with R-MAC feature vectors by weighted summation to generate the final feature descriptors. We have conducted experiments on public datasets and the results demonstrate that our proposed method could produce better retrieval performances than existing state-of-the-art algorithms. Further, higher performance could be achieved by performing these post-processing on the improved feature descriptors.https://www.mdpi.com/1099-4300/21/11/1037image retrievalpooling methodconvolutional neural networkfeature distribution entropy |
| spellingShingle | Pingping Liu Guixia Gou Huili Guo Danyang Zhang Hongwei Zhao Qiuzhan Zhou Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval Entropy image retrieval pooling method convolutional neural network feature distribution entropy |
| title | Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval |
| title_full | Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval |
| title_fullStr | Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval |
| title_full_unstemmed | Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval |
| title_short | Fusing Feature Distribution Entropy with R-MAC Features in Image Retrieval |
| title_sort | fusing feature distribution entropy with r mac features in image retrieval |
| topic | image retrieval pooling method convolutional neural network feature distribution entropy |
| url | https://www.mdpi.com/1099-4300/21/11/1037 |
| work_keys_str_mv | AT pingpingliu fusingfeaturedistributionentropywithrmacfeaturesinimageretrieval AT guixiagou fusingfeaturedistributionentropywithrmacfeaturesinimageretrieval AT huiliguo fusingfeaturedistributionentropywithrmacfeaturesinimageretrieval AT danyangzhang fusingfeaturedistributionentropywithrmacfeaturesinimageretrieval AT hongweizhao fusingfeaturedistributionentropywithrmacfeaturesinimageretrieval AT qiuzhanzhou fusingfeaturedistributionentropywithrmacfeaturesinimageretrieval |