遇见数据集

AxonData/age-estimation-minors-face-dataset

收藏
Hugging Face2026-04-03 更新2026-04-12 收录
官方服务:

资源简介:

--- license: cc-by-nc-4.0 task_categories: - image-classification tags: - age-estimation - age-detection - face-recognition - biometric - selfie - minors - age-verification - age-assurance - age-gating - age-prediction - COPPA pretty_name: "Face Age Detection: 10K Consented Selfies (10-30)" size_categories: - 10K<n<100K --- **10,000+ selfies | Verified age labels | 10-30 years | Multi-ethnic** ## The Problem New age-gating laws (COPPA 2.0, EU DSA, Australia Social Media Act) require apps to detect whether a user is under 18. Existing public face datasets don't solve this — UTKFace covers 0-116 with sparse teen coverage, IMDB-WIKI has noisy labels, MORPH is 80% one ethnicity ## The Solution This dataset targets the exact age range where models fail most: **10 to 30 years old**. Every image has a verified ground-truth age label, captured on real phones, with full legal consent including guardian signatures for minors ## Key Numbers | | | |---|---| | **Total images** | 10,000+ | | **Age range** | 10-30 (per-year folders) | | **Ethnicities** | Black, Caucasian, Hispanic | | **Capture device** | Smartphone front camera | | **Consent** | Full legal consent, guardian consent for minors | ![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F20109613%2F1911c8783f247fdcb115f56ceb1aef7c%2Fdataset_preview_age.png?generation=1775214156959812&alt=media) ## Labels `age` · `gender` · `ethnicity` ## Who Is This For - ML engineers building **under-18 age gates** for apps and platforms - Teams fine-tuning age estimation models on the **critical 13-25 segment** - Companies needing **consented, commercially licensable** face data - Researchers evaluating **demographic bias** in age prediction ## Full dataset with commercial license available at [Axon Labs](https://axonlab.ai)

--- 许可证:CC BY-NC 4.0 任务类别: - 图像分类(image-classification) 标签: - 年龄估计(age-estimation) - 年龄检测(age-detection) - 人脸识别(face-recognition) - 生物特征识别(biometric) - 自拍照(selfie) - 未成年人(minors) - 年龄验证(age-verification) - 年龄确认(age-assurance) - 年龄门控(age-gating) - 年龄预测(age-prediction) - COPPA(儿童在线隐私保护法,Children's Online Privacy Protection Act) 展示名称:"人脸年龄检测:10000张经同意的自拍照(10-30岁)" 样本量区间:10000 < 样本量 < 100000 --- **10000+张自拍照 | 经验证的年龄标注 | 年龄区间10-30岁 | 多族裔构成** ## 现存问题 新出台的年龄门控相关法案(如修订版《儿童在线隐私保护法》(COPPA 2.0)、欧盟数字服务法案(EU DSA)、澳大利亚《社交媒体法案》)均要求应用程序检测用户是否未满18岁。但现有公开人脸数据集均无法满足这一需求:UTKFace数据集覆盖年龄区间为0-116岁,但青少年样本分布稀疏;IMDB-WIKI数据集标注存在噪声;MORPH数据集样本中80%均为同一种族。 ## 解决方案 本数据集针对模型表现最差的精准年龄区间——**10至30岁**构建。所有图像均配有经核实的真实年龄标签,拍摄自真实智能手机,且已获得完整法律同意,未成年人样本还附带监护人签署的同意文件。 ## 核心统计数据 | | | |---|---| | **总图像数** | 10000+ | | **年龄区间** | 10-30岁(按年龄分文件夹存储) | | **族裔构成** | 黑人、白人、西班牙裔 | | **拍摄设备** | 智能手机前置摄像头 | | **知情同意** | 完整法律同意,未成年人样本附带监护人同意文件 | ![数据集预览图](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F20109613%2F1911c8783f247fdcb115f56ceb1aef7c%2Fdataset_preview_age.png?generation=1775214156959812&alt=media) ## 标注内容 `年龄` · `性别` · `族裔` ## 适用人群 - 为应用及平台开发**未满18岁年龄门控**的机器学习工程师 - 针对**13-25岁关键年龄区间**微调年龄估计模型的团队 - 需求**经同意、可商用授权**人脸数据的企业 - 评估年龄预测模型中**人口统计学偏差**的研究人员 ## 商用授权完整数据集 完整数据集(含商业授权)可在 [Axon Labs](https://axonlab.ai) 获取

提供机构:
AxonData
二维码
社区交流群
二维码
科研交流群
商业服务