APSIPA Transactions on Signal and Information Processing > Vol 12 > Issue 1

Deep Active Learning for Computer Vision: Past and Future

Rinyoichi Takezoe, Intellifusion Inc. and Peking University, China, Xu Liu, National University of Singapore, Singapore, Shunan Mao, Peking University, China, Marco Tianyu Chen, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences (SIAT), China, Zhanpeng Feng, Intellifusion Inc., China, Shiliang Zhang, Peking University, China, Xiaoyu Wang, Intellifusion Inc., China, fanghuaxue@gmail.com
 
Suggested Citation
Rinyoichi Takezoe, Xu Liu, Shunan Mao, Marco Tianyu Chen, Zhanpeng Feng, Shiliang Zhang and Xiaoyu Wang (2023), "Deep Active Learning for Computer Vision: Past and Future", APSIPA Transactions on Signal and Information Processing: Vol. 12: No. 1, e7. http://dx.doi.org/10.1561/116.00000057

Publication Date: 20 Mar 2023
© 2023 R. Takezoe, X. Liu, S. Mao, M. T. Chen, Z. Feng, S. Zhang and X. Wang
 
Subjects
 
Keywords
Computer visiondeep learningactive learning
 

Share

Open Access

This is published under the terms of CC BY-NC.

Downloaded: 2196 times

In this article:
Introduction 
Overview of Active Learning 
Vision Tasks Empowered by DeepAL 
Industrial Applications 
Discussion of Future Directions 
Conclusion 
References 

Abstract

As an important data selection schema, active learning emerges as the essential component when iterating an Artificial Intelligence (AI) model. It becomes even more critical given the dominance of deep neural network based models, which are composed of a large number of parameters and data hungry, in application. Despite its indispensable role for developing AI models, research on active learning is not as intensive as other research directions. In this paper, we present a review of active learning through deep active learning approaches from the following perspectives: (1) technical advancements in active learning, (2) applications of active learning in computer vision, (3) industrial systems leveraging or with potential to leverage active learning for data iteration, (4) current limitations and future research directions. We expect this paper to clarify the significance of active learning in a modern AI model manufacturing process and to bring additional research attention to active learning. By addressing data automation challenges and coping with automated machine learning systems, active learning will facilitate democratization of AI technologies by boosting model production at scale.

DOI:10.1561/116.00000057