Tsung-Yi Lin
Tsung-Yi Lin is a computer vision researcher at NVIDIA, and he co-authors the Microsoft COCO paper. The dataset holds 91 object types and 2.5 million labeled instances in 328,000 images. His papers include Focal Loss, which receives a best student paper award at ICCV 2017, and Simple Copy-Paste, a CVPR 2021 paper on data augmentation for instance segmentation. The COCO work receives the PAMI Mark Everingham Prize in 2023 and the Koenderink Prize in 2024. As of October 2026 he is a principal research scientist at NVIDIA Research.
Timeline
-
2014
The Microsoft COCO paper
In 2014 he co-writes the Microsoft COCO paper, published at ECCV. The dataset holds 91 object types and 2.5 million labeled instances in 328,000 images.
-
2017
Focal Loss at ICCV
In 2017 he co-writes Focal Loss for Dense Object Detection, published in the ICCV 2017 proceedings. The paper receives a best student paper award at ICCV 2017.
-
2021
Simple Copy-Paste at CVPR
In 2021 he co-writes Simple Copy-Paste, a CVPR paper on data augmentation for instance segmentation.
-
2023
The PAMI Mark Everingham Prize
In 2023 the COCO dataset receives the PAMI Mark Everingham Prize at ICCV.
-
2024
The Koenderink Prize
In 2024 ECCV gives its Koenderink Prize to the COCO paper, honoring his co-authorship.
-
As of October 2026
Principal research scientist at NVIDIA
He is a principal research scientist at NVIDIA Research, working on computer vision and machine learning.
Sources and credits
Photo: We have no published photo of Tsung-Yi Lin yet. Send us a free-licensed photo.
- Microsoft COCO: Common Objects in Context - Microsoft Research
- [1405.0312] Microsoft COCO: Common Objects in Context
- ICCV 2017 Open Access Repository
- Tsung-Yi Lin
- CVPR 2021 Open Access Repository
- NVIDIA unveils AI breakthroughs at European vision conference - CO/AI
- Author: Tsung-Yi Lin | NVIDIA Technical Blog