| Title: |
Families in the Wild (FIW): Large-Scale Kinship Image Database and Benchmarks |
| Article URLs: |
|
| Alternative Article URLs: |
https://www.dropbox.com/s/au1fv7f8m4bq2ab/acm-mm-short.pdf?dl=0 |
| Authors: |
Joseph P. Robinson |
-
Northeastern University, Department of Electrical and Computer Engineering
|
| Ming Shao |
-
Northeastern University, Department of Electrical and Computer Engineering
|
| Yue Wu |
-
Northeastern University, Department of Electrical and Computer Engineering
|
| Yun Fu |
-
Northeastern University, Department of Electrical and Computer Engineering
-
Northeastern University, College of Computer and Information Science
|
| Sharing: |
Research produced artifacts
|
| Verification: |
Authors have
verified
information
|
| Artifact Evaluation Badge: |
none
|
| Artifact URLs: |
|
| Artifact Correspondence Email Addresses: |
|
| NSF Award Numbers: |
|
| DBLP Key: |
conf/mm/RobinsonSWF16
|
| Author Comments: |
Families In the Wild (FIW) is a large-scale image database for automatic kinship recognition based tasks. With over 12,000 family photos of 1,000 families FIW provides the data required for many modern-day deep learning approaches and it is, by far, the largest resource of its type to date. We benchmark two tasks using FIW: (1) kinship verification (one-to-one), which spans 11 relationship types and about 650,000 pairs; (2) family classification (one-to-many), which uses 524 families consisted of 12,007 facial images. We expect additional tasks and use-cases of FIW to be put forward through creativity and further research in due time-- but only by us, but other researchers (from academia and industry), and a combination of both. FIW is a continuing effort, as we plan for the database to continually grow in terms of size and usefulness as time transpires. |