• Publications
  • Influence
Squeeze-and-Excitation Networks
TLDR
This work proposes a novel architectural unit, which is term the “Squeeze-and-Excitation” (SE) block, that adaptively recalibrates channel-wise feature responses by explicitly modelling interdependencies between channels and shows that these blocks can be stacked together to form SENet architectures that generalise extremely effectively across different datasets.
Constant Time Weighted Median Filtering for Stereo Matching and Beyond
TLDR
It is discovered that with this refinement, even the simple box filter aggregation achieves comparable accuracy with various sophisticated aggregation methods (with the same refinement), revealing that the previously overlooked refinement can be at least as crucial as aggregation.
Transformer in Transformer
TLDR
It is pointed out that the attention inside these local patches are also essential for building visual transformers with high performance and a new architecture, namely, Transformer iN Transformer (TNT), is explored.
A Closed-Form Solution to Retinex with Nonlocal Texture Constraints
TLDR
This work formulate the decomposition problem as the minimization of a quadratic function which incorporates both the retinex constraint and the nonlocal texture constraint, and can be solved in closed form with the standard conjugate gradient algorithm.
Handling motion blur in multi-frame super-resolution
TLDR
The method proposed in this paper tackles the issue of ubiquitous motion blur easily fails multi-frame super-resolution by optimally searching least blurred pixels in MFSR and produces sharp and higher-resolution results given input of challenging low-resolution noisy and blurred sequences.
FreePipe: a programmable parallel rendering architecture for efficient multi-fragment effects
TLDR
FreePipe is presented, a system for programmable parallel rendering that can run entirely on current graphics hardware and has performance comparable with the traditional graphics pipeline.
Efficient depth peeling via bucket sort
TLDR
An efficient algorithm for multi-layer depth peeling via bucket sort of fragments on GPU, which makes it possible to capture up to 32 layers simultaneously with correct depth ordering in a single geometry pass, and is free of read-modify-write (RMW) hazards.
Adaptive multi-task learning for fine-grained categorization
TLDR
This paper proposes a novel multi-task learning method to adaptively share information that captures the relationships among tasks and identifies the disparities of each task simultaneously, thus can flexibly exploit the shared information.
Real-time 3D fluid simulation on GPU with complex obstacles
TLDR
The method generates the boundary from a 3D scene with solid clipping, making the computation run on GPU despite of the complexity of the whole geometry scene, and modify the values on the boundaries according to the boundary conditions.
Photorealistic rendering of knitwear using the lumislice
TLDR
A method for efficient synthesis of photorealistic free-form knitwear that accommodates varying levels of detail and capitalizes on hardware-assisted transparency blending and a technique for generating soft shadows from yarn is introduced.
...
1
2
3
4
5
...