Anastasis Stathopoulos
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Anastasis Stathopoulos has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Long Zhao
Google Research CA USA
2 shared publications - Manmohan Chandraker
University of California San Diego La Jolla 92093 CA USA
2 shared publications - Samuel Schulter
NEC Labs America
2 shared publications - B. G. Vijay Kumar
NEC Labs America
1 shared publication - Dimitrios Metaxas
Department of Physics National Technical University of Athens Zografou Campus GR 15773 Athens Greece
1 shared publication Dimitris Metaxas
1 shared publication
Latest Publications
- 6 November 2022
- Published by Springer Nature
- p. 159-175
Exploiting Unlabeled Data with Vision and Language Models for Object Detection
- 6 November 2022
- Published by Springer Nature
- p. 159-175
Exploiting Unlabeled Data with Vision and Language Models for Object Detection
- Open Access
Exploiting Unlabeled Data with Vision and Language Models for Object Detection
Building robust and generic object detection frameworks requires scaling to larger label spaces and bigger training datasets. However, it is prohibitively costly to acquire annotations for thousands of categories at a large scale. We propose a novel method that leverages the rich semantics available in recent vision and language models to localize and classify objects in unlabeled images, effectively generating pseudo labels for object detection. Starting with a generic and class-agnostic region proposal mechanism, we use vision and language models to categorize each region of an image into any object category that is required for downstream tasks. We demonstrate the value of the generated pseudo labels in two specific tasks, open-vocabulary detection, where a model needs to generalize to unseen object categories, and semi-supervised object detection, where additional unlabeled images can be used to improve the model. Our empirical evaluation shows the effectiveness of the pseudo labels in both tasks, where we outperform competitive baselines and achieve a novel state-of-the-art for open-vocabulary object detection. Our code is available at https://github.com/xiaofeng94/VL-PLM.
Open AccessExploiting Unlabeled Data with Vision and Language Models for Object Detection
Building robust and generic object detection frameworks requires scaling to larger label spaces and bigger training datasets. However, it is prohibitively costly to acquire annotations for thousands of categories at a large scale. We propose a novel method that leverages the rich semantics available in recent vision and language models to localize and classify objects in unlabeled images, effectively generating pseudo labels for object detection. Starting with a generic and class-agnostic region proposal mechanism, we use vision and language models to categorize each region of an image into any object category that is required for downstream tasks. We demonstrate the value of the generated pseudo labels in two specific tasks, open-vocabulary detection, where a model needs to generalize to unseen object categories, and semi-supervised object detection, where additional unlabeled images can be used to improve the model. Our empirical evaluation shows the effectiveness of the pseudo labels in both tasks, where we outperform competitive baselines and achieve a novel state-of-the-art for open-vocabulary object detection. Our code is available at https://github.com/xiaofeng94/VL-PLM.