Multicolumn networks for face recognition
<p>The objective of this work is set-based face recognition, i.e. to decide if two sets of images of a face are of the same person or not. Conventionally, the set-wise feature descriptor is computed as an average of the descriptors from individual face images within the set. In this paper, we...
| Autors principals: | , |
|---|---|
| Format: | Conference item |
| Idioma: | English |
| Publicat: |
British Machine Vision Association
2018
|
| _version_ | 1826313317705056256 |
|---|---|
| author | Xie, W Zisserman, A |
| author_facet | Xie, W Zisserman, A |
| author_sort | Xie, W |
| collection | OXFORD |
| description | <p>The objective of this work is set-based face recognition, i.e. to decide if two sets of images of a face are of the same person or not. Conventionally, the set-wise feature descriptor is computed as an average of the descriptors from individual face images within the set. In this paper, we design a neural network architecture that learns to aggregate based on both “visual” quality (resolution, illumination), and “content” quality (relative importance for discriminative classification).</p> <br/>
<p>To this end, we propose a Multicolumn Network (MN) that takes a set of images (the number in the set can vary) as input, and learns to compute a fix-sized feature descriptor for the entire set. To encourage high-quality representations, each individual input image is first weighted by its “visual” quality, determined by a self-quality assessment module, and followed by a dynamic recalibration based on “content” qualities relative to the other images within the set. Both of these qualities are learnt implicitly during training for setwise classification. Comparing with the previous state-of-the-art architectures trained with the same dataset (VGGFace2), our Multicolumn Networks show an improvement of between 2-6% on the IARPA IJB face recognition benchmarks, and exceed the state of the art for all methods on these benchmarks.</p> |
| first_indexed | 2024-03-07T04:20:43Z |
| format | Conference item |
| id | oxford-uuid:caea1912-d942-4b4d-9e3a-99e251db6e40 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-09-25T04:11:07Z |
| publishDate | 2018 |
| publisher | British Machine Vision Association |
| record_format | dspace |
| spelling | oxford-uuid:caea1912-d942-4b4d-9e3a-99e251db6e402024-06-17T14:32:28ZMulticolumn networks for face recognitionConference itemhttp://purl.org/coar/resource_type/c_5794uuid:caea1912-d942-4b4d-9e3a-99e251db6e40EnglishSymplectic Elements at OxfordBritish Machine Vision Association2018Xie, WZisserman, A<p>The objective of this work is set-based face recognition, i.e. to decide if two sets of images of a face are of the same person or not. Conventionally, the set-wise feature descriptor is computed as an average of the descriptors from individual face images within the set. In this paper, we design a neural network architecture that learns to aggregate based on both “visual” quality (resolution, illumination), and “content” quality (relative importance for discriminative classification).</p> <br/> <p>To this end, we propose a Multicolumn Network (MN) that takes a set of images (the number in the set can vary) as input, and learns to compute a fix-sized feature descriptor for the entire set. To encourage high-quality representations, each individual input image is first weighted by its “visual” quality, determined by a self-quality assessment module, and followed by a dynamic recalibration based on “content” qualities relative to the other images within the set. Both of these qualities are learnt implicitly during training for setwise classification. Comparing with the previous state-of-the-art architectures trained with the same dataset (VGGFace2), our Multicolumn Networks show an improvement of between 2-6% on the IARPA IJB face recognition benchmarks, and exceed the state of the art for all methods on these benchmarks.</p> |
| spellingShingle | Xie, W Zisserman, A Multicolumn networks for face recognition |
| title | Multicolumn networks for face recognition |
| title_full | Multicolumn networks for face recognition |
| title_fullStr | Multicolumn networks for face recognition |
| title_full_unstemmed | Multicolumn networks for face recognition |
| title_short | Multicolumn networks for face recognition |
| title_sort | multicolumn networks for face recognition |
| work_keys_str_mv | AT xiew multicolumnnetworksforfacerecognition AT zissermana multicolumnnetworksforfacerecognition |