Rainbow Group Seminars

Modern XR devices can achieve high frame rate, high resolution, wide color gamut, and brighter displays in order to achieve perceptual realism. However, these requirements lead to significant challenges in terms of battery life, with current commercial devices only lasting up to 3 hours on a single charge. Many techniques...

Modern XR devices can achieve high frame rate, high resolution, wide color gamut, and brighter displays in order to achieve perceptual realism. However, these requirements lead to significant challenges in terms of battery life, with current commercial devices only lasting up to 3 hours on a single charge. Many techniques...

As a form of computer graphics, vector graphics represents visual content based directly on geometric shapes (via command lines and arguments), and is widely used in scientific and artistic applications, including architecture, surveying, 3D rendering, typography, and graphic design. Compared with raster graphics, vector...

Recent claims about the superiority of deep-nets to model human vision are based on the reproduction of either (a) neural recordings from different brain layers or (b) high-level behaviors included in Brain-Score. However, high correlations in such tasks do not guarantee the functional similarity of the underlying...

The empirical theory of vision suggests that what we see reflects statistical regularities learned through experience rather than direct representations of physical reality. Recent advances in machine learning, particularly in foundation models, have revealed intriguing parallels with human visual perception. However, the...

Angular information is applied across domains such as navigation, virtual and augmented reality, and robotics to tackle research problems related to directions and orientations. It can be encoded into polarization to reveal hidden content, and it can also be encoded into parallax to track an object with high accuracy. This...

To better train and evaluate 3D reconstruction methods (NeRF, Gaussian Splatting) or 3D generative models, both for static (3D) and dynamic (4D) scenes, we will develop a new full-reference quality metric and no-reference loss function. Those will be trained and validated on a new 4D quality dataset, with the subjective...

Streaming rendered content is an attractive way to bring high-quality graphics to billions of mobile devices that do not have sufficient rendering power. Existing solutions render content on a server at a fixed frame rate, typically 30 or 60 frames per second, and reduce resolution when bandwidth is restricted. Here, we...

Hanxue Liang will present his recent work BulletTimer. BulletTimer is the first motion-aware feed-forward model for real-time dynamic scene reconstruction. It takes a monocular video as input and reconstructs a 3D Gaussian Splatting (3DGS) representation at any desired timestamp in a feed-forward manner within just 150 ms...

Vision, as a symphony of physics and intelligence, encompasses two interconnected dimensions: sensing and perception. Visual sensing, grounded in physical optics, harnesses computational intelligence to enhance light capture. Visual perception, driven by intelligence, is now empowered by physical optics to interpret visual...