top of page



Comparison and Prospects of the Performance of Machine Vision and Human Vision in Age Estimation


Project Details 計畫內容概述

Face recognition, gender recognition, age estimation, and expression recognition are commonly used to analyze face-related information. Among them, age estimation is the most difficult to utilize human's own vision. We can often easily determine the gender, expression, or whether a person is an acquaintance from a face, but we need to know the age of people we often meet by asking them, and we dare not easily check it visually, which indicates that it may be more difficult to estimate the age of a person by using a human face than by analyzing other face information. Although it is not easy to determine the exact age of a person with the naked eye, we can often point out a fairly accurate age range, so if we can learn more about how human vision determines and intercepts the age range, we may be able to improve the age estimation using machine vision. The following points are planned for this study:

  1. Create an accurate age estimator with machine vision, which is as good as the latest age estimators;

  2. To design experiments to understand how human vision determines age and intercepts the age interval, and to explore how human vision can be utilized to improve the age estimation by machine vision;
  3. To combine the individual strengths of the Taiwan and UK teams to investigate the effects of machine vision, human vision and wrinkle detection on age estimation and to improve the estimation performance;

  4. To develop research projects with more application value and development potential to facilitate longer-term research and collaboration between the two organizations.


  1. 以機器視覺製作一個效能精確的年齡估測器,其效能不亞於目前最新的年齡估測器;
  2. ​設計實驗了解人類視覺如何判斷年齡與截取年齡區間,並探討如何利用人類視覺改善機器視覺的年齡估測;
  3. 結合台灣和英國團隊個別的優勢,探討機器視覺、人類視覺和皺紋偵測對年齡估測的影響,改進估測效能;
  4. 擬定更具應用價值和發展潛力的研究項目,以利雙方更長遠的研究與合作。




Project No.


Project Partners

  1. National Taiwan University of Science and Technology (NTUST) in Taiwan

  2. Manchester Metropolitan University

Principal Investigators

  1. Professor Gee Sern Hsu (Taiwan)

  2. Professor Moi Hoon Yap (UK)

Contact 聯絡資訊

Publication 期刊論文發表

  • Jireh Jam, Connah Kendrick, Kevin Walker, Vincent Drouard, Gee-Sern Hsu, Moi Hoon Yap, A comprehensive review of past and present image inpainting methods, Comput. Vis. Image Underst. (2021)

  • *Gee-Sern Hsu, Wen-Fong Huang, Moi Hoon Yap. Edge-Embedded Multi-Dropout Framework for Real-Time Face Alignment, IEEE Access (2020)

  • Choon-Ching Ng, Moi Hoon Yap, Yi-Tseng Cheng, Gee-Sern Hsu. Hybrid Ageing Patterns for Face Age Estimation. Image Vis. Comput. (2017)

bottom of page