Task | Accuracy (%) |
---|
ID | Gender | ID | Gender |
---|
O | Â | 98.28 | - |
 | O | - | 97.70 |
O | O | 98.97 | 96.55 |
- When dealing with multiple tasks, it is more accurate in identification task. Since the model trains for both identification and gender classification task, the process to classify gender can assists in identifying human being. For gender classification task, our multitask model seems a bit less accurate than that in single task model. However, considering the number of parameters for models, the proposed model is efficient