S
Seong Joon Oh
Researcher at Naver Corporation
Publications - 65
Citations - 6472
Seong Joon Oh is an academic researcher from Naver Corporation. The author has contributed to research in topics: Computer science & Metric (mathematics). The author has an hindex of 24, co-authored 52 publications receiving 2971 citations. Previous affiliations of Seong Joon Oh include Max Planck Society & Kyoto University.
Papers
More filters
Proceedings ArticleDOI
CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features
TL;DR: CutMix as discussed by the authors augments the training data by cutting and pasting patches among training images, where the ground truth labels are also mixed proportionally to the area of the patches.
Proceedings ArticleDOI
What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis
Jeonghun Baek,Geewook Kim,Junyeop Lee,Sungrae Park,Dongyoon Han,Sangdoo Yun,Seong Joon Oh,Hwalsuk Lee +7 more
TL;DR: In this paper, a unified four-stage scene text recognition (STR) framework is introduced to compare the performance of different models. But, the performance gap results from inconsistencies in the training and evaluation datasets.
Posted Content
Towards Reverse-Engineering Black-Box Neural Networks
TL;DR: This paper showed that the revealed internal information helps generate more effective adversarial examples against the black-box model, which can be used for better protection of private content from automatic recognition models using adversarial example.
Posted Content
CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
TL;DR: Patches are cut and pasted among training images where the ground truth labels are also mixed proportionally to the area of the patches, and CutMix consistently outperforms state-of-the-art augmentation strategies on CIFAR and ImageNet classification tasks, as well as on ImageNet weakly-supervised localization task.
Proceedings ArticleDOI
Exploiting Saliency for Object Segmentation from Image Level Labels
TL;DR: In this article, a saliency model was proposed to exploit prior knowledge on the object extent and image statistics to recover 80% of the fully-supervised performance in weakly supervised pixel-wise semantic labeling.