Wonjun Yoon

Hoon Kim, Minje Jang, Wonjun Yoon et al.

We introduce a co-designed approach for human portrait relighting that combines a physics-guided architecture with a pre-training framework. Drawing on the Cook-Torrance reflectance model, we have meticulously configured the architecture design to precisely simulate light-surface interactions. Furthermore, to overcome the limitation of scarce high-quality lightstage data, we have developed a self-supervised pre-training strategy. This novel combination of accurate physical modeling and expanded training dataset establishes a new benchmark in relighting realism.

Geondo Park, Chihye Han, Wonjun Yoon et al.

Visual-semantic embedding enables various tasks such as image-text retrieval, image captioning, and visual question answering. The key to successful visual-semantic embedding is to express visual and textual data properly by accounting for their intricate relationship. While previous studies have achieved much advance by encoding the visual and textual data into a joint space where similar concepts are closely located, they often represent data by a single vector ignoring the presence of multiple important components in an image or text. Thus, in addition to the joint embedding space, we propose a novel multi-head self-attention network to capture various components of visual and textual data by attending to important parts in data. Our approach achieves the new state-of-the-art results in image-text retrieval tasks on MS-COCO and Flicker30K datasets. Through the visualization of the attention maps that capture distinct semantic components at multiple positions in the image and the text, we demonstrate that our method achieves an effective and interpretable visual-semantic joint space.

Hyeonseob Nam, HyunJae Lee, Jongchan Park et al.

Convolutional Neural Networks (CNNs) often fail to maintain their performance when they confront new test domains, which is known as the problem of domain shift. Recent studies suggest that one of the main causes of this problem is CNNs' strong inductive bias towards image styles (i.e. textures) which are sensitive to domain changes, rather than contents (i.e. shapes). Inspired by this, we propose to reduce the intrinsic style bias of CNNs to close the gap between domains. Our Style-Agnostic Networks (SagNets) disentangle style encodings from class categories to prevent style biased predictions and focus more on the contents. Extensive experiments show that our method effectively reduces the style bias and makes the model more robust under domain shift. It achieves remarkable performance improvements in a wide range of cross-domain tasks including domain generalization, unsupervised domain adaptation, and semi-supervised domain adaptation on multiple datasets.

Chihye Han, Wonjun Yoon, Gihyun Kwon et al.

The recent success of brain-inspired deep neural networks (DNNs) in solving complex, high-level visual tasks has led to rising expectations for their potential to match the human visual system. However, DNNs exhibit idiosyncrasies that suggest their visual representation and processing might be substantially different from human vision. One limitation of DNNs is that they are vulnerable to adversarial examples, input images on which subtle, carefully designed noises are added to fool a machine classifier. The robustness of the human visual system against adversarial examples is potentially of great importance as it could uncover a key mechanistic feature that machine vision is yet to incorporate. In this study, we compare the visual representations of white- and black-box adversarial examples in DNNs and humans by leveraging functional magnetic resonance imaging (fMRI). We find a small but significant difference in representation patterns for different (i.e. white- versus black- box) types of adversarial examples for both humans and DNNs. However, human performance on categorical judgment is not degraded by noise regardless of the type unlike DNN. These results suggest that adversarial examples may be differentially represented in the human visual system, but unable to affect the perceptual experience.

Wonjun Yoon, Sol-A Kim, Jaesik Choi

With the advance in robotic hardware and intelligent software, humanoid robot plays an important role in various tasks including service for human assistance and heavy job for hazardous industry. Recent advances in task learning enable humanoid robots to conduct dexterous manipulation tasks such as grasping objects and assembling parts of furniture. Operating objects without physical movements is an even more challenging task for humanoid robot because effects of actions may not be clearly seen in the physical configuration space and meaningful actions could be very complex in a long time horizon. As an example, playing a mobile game in a smart device has such challenges because both swipe actions and complex state transitions inside the smart devices in a long time horizon. In this paper, we solve this problem by introducing an integrated architecture which connects end-to-end dataflow from sensors to actuators in a humanoid robot to operate smart devices. We implement our integrated architecture in the Baxter Research Robot and experimentally demonstrate that the robot with our architecture could play a challenging mobile game, the 2048 game, as accurate as in a simulated environment.