Age, Gender & Emotion Benchmarks
Real Age Estimation
This table shows the mean absolute error for our methods along with competitive approaches. Our method outperforms the second best method by 1.58 years.
| Method | MAE |
| Sighthound | 5.76 |
| Rothe et al. 1 | 7.34 |
| Microsoft 2 | 7.62 |
| Face++ 3 | 11.04 |
Age Classification Accuracy of Group Dataset
We report both the exact accuracy as well as the 1-off accuracy.
| Method | top1 | 1-off |
| Sighthound | 70.5% | 96.2% |
| Hou et al. 4 | 65.0% | 96.1% |
| Rothe et al. 1 | 62.3% | 94.3% |
| Dong et al. 5 | 56.0% | 92.0% |
| Gallegher et al. 6 | 42.9% | 78.1% |
Emotion Recognition Accuracy on Sighthound Dataset
| Method | Accuracy |
| Sighthound | 76.1% |
| Microsoft 2 | 61.3% |
Gender Recognition Accuracy on the Adience Benchmark
| Method | Accuracy |
| Sighthound | 91.00% |
| Microsoft 2 | 90.86% |
| Rothe et al. 1 | 88.75% |
| Levi and Hassner 7 | 86.80% |
| Kairos 8 | 84.66% |
| Face++ 3 | 83.04% |
References
1. Rothe, R., Radu Timofte, L.V.G.: Dex: Deep expectation of apparent age from asingle image. In: International Conference on Computer Vision (ICCV). (2015)
2. Microsoft-Face-API.: (https://www.microsoft.com/cognitive-services/en-us/face-api.)
3. Face++.: (http://old.faceplusplus.com/demo-detect/)
4. Hou, L., Yu, C.P., Samaras., D.: Squared earth mover’s distance-based loss fortraining deep neural networks. In: arXiv. (2016)
5. Yuan, D., Liu, Y., Lian., S.: Automatic age estimation based on deep learningalgorithm. In: Neurocomputing. (2016)
6. Gallagher, A.C., Chen., T.: Understanding images of groups of people. In: CVPR.(2009)
7. Levi, G., Hassner., T.: Age and gender classification using convolutional neural networks. In: CVPRW. (2015)
8. Kairos.: (https://www.kairos.com/kairos-2.0/demos)