Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition
The key to fine-grained aircraft recognition is discovering the subtle traits that can distinguish different subcategories. Early approaches leverage part annotations of fine-grained objects to derive rich representations. However, manual labeling part information is cumbersome. In response to this...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-12-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/24/5132 |
| _version_ | 1827669826988933120 |
|---|---|
| author | Xiaolan Huang Kai Xu Chuming Huang Chengrui Wang Kun Qin |
| author_facet | Xiaolan Huang Kai Xu Chuming Huang Chengrui Wang Kun Qin |
| author_sort | Xiaolan Huang |
| collection | DOAJ |
| description | The key to fine-grained aircraft recognition is discovering the subtle traits that can distinguish different subcategories. Early approaches leverage part annotations of fine-grained objects to derive rich representations. However, manual labeling part information is cumbersome. In response to this issue, previous CNN-based methods reuse the backbone network to extract part-discrimination features, the inference process of which consumes much time. Therefore, we introduce generalized multiple instance learning (MIL) into fine-grained recognition. In generalized MIL, an aircraft is assumed to consist of multiple instances (such as head, tail, and body). Firstly, instance-level representations are obtained by the feature extractor and instance conversion component. Secondly, the obtained instance features are scored by an MIL classifier, which can yield high-level part semantics. Finally, a fine-grained object label is inferred by a MIL pooling function that aggregates multiple instance scores. The proposed approach is trained end-to-end without part annotations and complex location networks. Experimental evidence is conducted to prove the feasibility and effectiveness of our approach on combined aircraft images (CAIs). |
| first_indexed | 2024-03-10T03:11:30Z |
| format | Article |
| id | doaj.art-5560279c85a34eb6adb1aa3714a6e5b5 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T03:11:30Z |
| publishDate | 2021-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-5560279c85a34eb6adb1aa3714a6e5b52023-11-23T10:25:15ZengMDPI AGRemote Sensing2072-42922021-12-011324513210.3390/rs13245132Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft RecognitionXiaolan Huang0Kai Xu1Chuming Huang2Chengrui Wang3Kun Qin4School of Geography and Information Engineering, China University of Geosciences, Wuhan 430078, ChinaSchool of Geography and Information Engineering, China University of Geosciences, Wuhan 430078, ChinaSchool of Geography and Information Engineering, China University of Geosciences, Wuhan 430078, ChinaSchool of Geography and Information Engineering, China University of Geosciences, Wuhan 430078, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, ChinaThe key to fine-grained aircraft recognition is discovering the subtle traits that can distinguish different subcategories. Early approaches leverage part annotations of fine-grained objects to derive rich representations. However, manual labeling part information is cumbersome. In response to this issue, previous CNN-based methods reuse the backbone network to extract part-discrimination features, the inference process of which consumes much time. Therefore, we introduce generalized multiple instance learning (MIL) into fine-grained recognition. In generalized MIL, an aircraft is assumed to consist of multiple instances (such as head, tail, and body). Firstly, instance-level representations are obtained by the feature extractor and instance conversion component. Secondly, the obtained instance features are scored by an MIL classifier, which can yield high-level part semantics. Finally, a fine-grained object label is inferred by a MIL pooling function that aggregates multiple instance scores. The proposed approach is trained end-to-end without part annotations and complex location networks. Experimental evidence is conducted to prove the feasibility and effectiveness of our approach on combined aircraft images (CAIs).https://www.mdpi.com/2072-4292/13/24/5132fine-grained image recognitionaircraft recognitionmultiple instance learningloss function |
| spellingShingle | Xiaolan Huang Kai Xu Chuming Huang Chengrui Wang Kun Qin Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition Remote Sensing fine-grained image recognition aircraft recognition multiple instance learning loss function |
| title | Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition |
| title_full | Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition |
| title_fullStr | Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition |
| title_full_unstemmed | Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition |
| title_short | Multiple Instance Learning Convolutional Neural Networks for Fine-Grained Aircraft Recognition |
| title_sort | multiple instance learning convolutional neural networks for fine grained aircraft recognition |
| topic | fine-grained image recognition aircraft recognition multiple instance learning loss function |
| url | https://www.mdpi.com/2072-4292/13/24/5132 |
| work_keys_str_mv | AT xiaolanhuang multipleinstancelearningconvolutionalneuralnetworksforfinegrainedaircraftrecognition AT kaixu multipleinstancelearningconvolutionalneuralnetworksforfinegrainedaircraftrecognition AT chuminghuang multipleinstancelearningconvolutionalneuralnetworksforfinegrainedaircraftrecognition AT chengruiwang multipleinstancelearningconvolutionalneuralnetworksforfinegrainedaircraftrecognition AT kunqin multipleinstancelearningconvolutionalneuralnetworksforfinegrainedaircraftrecognition |