Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval
Along with a new volume of images containing valuable information about our past, the digitization of historical territorial imagery has brought the challenge of understanding and interconnecting collections with unique or rare representation characteristics, and sparse metadata. Content-based image...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-08-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/16/3080 |
| _version_ | 1797522199328849920 |
|---|---|
| author | Dimitri Gominski Valérie Gouet-Brunet Liming Chen |
| author_facet | Dimitri Gominski Valérie Gouet-Brunet Liming Chen |
| author_sort | Dimitri Gominski |
| collection | DOAJ |
| description | Along with a new volume of images containing valuable information about our past, the digitization of historical territorial imagery has brought the challenge of understanding and interconnecting collections with unique or rare representation characteristics, and sparse metadata. Content-based image retrieval offers a promising solution in this context, by building links in the data without relying on human supervision. However, while the latest propositions in deep learning have shown impressive results in applications linked to feature learning, they often rely on the hypothesis that there exists a training dataset matching the use case. Increasing generalization and robustness to variations remains an open challenge, poorly understood in the context of real-world applications. Introducing the <span style="font-variant: small-caps;">alegoria</span> benchmark, containing multi-date vertical and oblique aerial digitized photography mixed with more modern street-level pictures, we formulate the problem of low-data, heterogeneous image retrieval, and propose associated evaluation setups and measures. We propose a review of ideas and methods to tackle this problem, extensively compare state-of-the-art descriptors and propose a new multi-descriptor diffusion method to exploit their comparative strengths. Our experiments highlight the benefits of combining descriptors and the compromise between absolute and cross-domain performance. |
| first_indexed | 2024-03-10T08:26:06Z |
| format | Article |
| id | doaj.art-21de93e8a14f4203925e93da3f61fd12 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T08:26:06Z |
| publishDate | 2021-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-21de93e8a14f4203925e93da3f61fd122023-11-22T09:31:29ZengMDPI AGRemote Sensing2072-42922021-08-011316308010.3390/rs13163080Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance RetrievalDimitri Gominski0Valérie Gouet-Brunet1Liming Chen2LaSTIG, IGN-ENSG, Gustave Eiffel University, 77420 Champs-sur-Marne, FranceLaSTIG, IGN-ENSG, Gustave Eiffel University, 77420 Champs-sur-Marne, FranceLIRIS, École Centrale de Lyon, 69134 Écully, FranceAlong with a new volume of images containing valuable information about our past, the digitization of historical territorial imagery has brought the challenge of understanding and interconnecting collections with unique or rare representation characteristics, and sparse metadata. Content-based image retrieval offers a promising solution in this context, by building links in the data without relying on human supervision. However, while the latest propositions in deep learning have shown impressive results in applications linked to feature learning, they often rely on the hypothesis that there exists a training dataset matching the use case. Increasing generalization and robustness to variations remains an open challenge, poorly understood in the context of real-world applications. Introducing the <span style="font-variant: small-caps;">alegoria</span> benchmark, containing multi-date vertical and oblique aerial digitized photography mixed with more modern street-level pictures, we formulate the problem of low-data, heterogeneous image retrieval, and propose associated evaluation setups and measures. We propose a review of ideas and methods to tackle this problem, extensively compare state-of-the-art descriptors and propose a new multi-descriptor diffusion method to exploit their comparative strengths. Our experiments highlight the benefits of combining descriptors and the compromise between absolute and cross-domain performance.https://www.mdpi.com/2072-4292/13/16/3080CBIRcross-domaincultural heritagebenchmarkingdiffusion |
| spellingShingle | Dimitri Gominski Valérie Gouet-Brunet Liming Chen Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval Remote Sensing CBIR cross-domain cultural heritage benchmarking diffusion |
| title | Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval |
| title_full | Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval |
| title_fullStr | Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval |
| title_full_unstemmed | Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval |
| title_short | Connecting Images through Sources: Exploring Low-Data, Heterogeneous Instance Retrieval |
| title_sort | connecting images through sources exploring low data heterogeneous instance retrieval |
| topic | CBIR cross-domain cultural heritage benchmarking diffusion |
| url | https://www.mdpi.com/2072-4292/13/16/3080 |
| work_keys_str_mv | AT dimitrigominski connectingimagesthroughsourcesexploringlowdataheterogeneousinstanceretrieval AT valeriegouetbrunet connectingimagesthroughsourcesexploringlowdataheterogeneousinstanceretrieval AT limingchen connectingimagesthroughsourcesexploringlowdataheterogeneousinstanceretrieval |