Face Age and Gender Estimation System Development

Face Age and Gender Estimation System Development

Age and gender estimation from facial images is a computer vision task with applications in retail analytics (visitor demographic profile), adaptive content systems, medical research, age verification (age gate). Both tasks are often implemented by a single multitask model.

Multitask Architecture

import torch
import torch.nn as nn
import timm

class AgeGenderModel(nn.Module):
    """Single model for simultaneous age and gender prediction"""
    def __init__(self, pretrained_backbone: str = 'efficientnet_b2'):
        super().__init__()
        backbone = timm.create_model(pretrained_backbone, pretrained=True, num_classes=0)
        self.backbone = backbone
        feat_dim = backbone.num_features  # 1408 for B2

        # Shared representation
        self.shared = nn.Sequential(
            nn.Linear(feat_dim, 512),
            nn.GELU(),
            nn.Dropout(0.3)
        )

        # Separate heads for each task
        self.age_head = nn.Linear(512, 1)      # regression (MAE)
        self.gender_head = nn.Linear(512, 2)   # classification (CE)

    def forward(self, x):
        features = self.backbone(x)
        shared = self.shared(features)
        age = self.age_head(shared).squeeze()
        gender_logits = self.gender_head(shared)
        return age, gender_logits

Age as regression vs classification: regression gives continuous result (32.4 years), classification by ranges (30–35 years) is less accurate but more convenient for some applications. Distributional regression (DLDL — Distribution Learning) is the best approach: age is modeled as probability distribution rather than point value.

Datasets

Preprocessing and Augmentation

import albumentations as A
from albumentations.pytorch import ToTensorV2

train_transform = A.Compose([
    A.Resize(224, 224),
    A.HorizontalFlip(p=0.5),
    A.RandomBrightnessContrast(brightness_limit=0.3, contrast_limit=0.3, p=0.5),
    A.GaussianBlur(blur_limit=(3, 7), p=0.2),
    A.CoarseDropout(max_holes=4, max_height=30, max_width=30, p=0.3),
    A.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
    ToTensorV2()
])

Loss Functions for Multitask Learning

def multitask_loss(age_pred, age_true, gender_logits, gender_true,
                   age_weight=1.0, gender_weight=0.5):
    # MAE for age + CE for gender
    age_loss = nn.L1Loss()(age_pred, age_true.float())
    gender_loss = nn.CrossEntropyLoss()(gender_logits, gender_true)

    # Uncertainty weighting (Kendall et al.)
    return age_weight * age_loss + gender_weight * gender_loss

Performance Metrics

MAE 4–6 years is typical for "in-the-wild" (selfies, varying quality photos). In controlled conditions (frontal portrait, good lighting) — 3–4 years.

Ethics and Bias

Models trained on IMDB-Wiki have underrepresentation of elderly people and some ethnic groups. FairFace dataset is specially balanced to reduce bias. When using for decision-making (age gate) — mandatory fairness testing across demographic groups.