T
Tyng-Luh Liu
Researcher at Academia Sinica
Publications - 104
Citations - 3873
Tyng-Luh Liu is an academic researcher from Academia Sinica. The author has contributed to research in topics: Computer science & Image segmentation. The author has an hindex of 27, co-authored 94 publications receiving 3432 citations. Previous affiliations of Tyng-Luh Liu include Courant Institute of Mathematical Sciences & New York University.
Papers
More filters
Proceedings ArticleDOI
Local discriminant embedding and its variants
TL;DR: A new approach, called local discriminant embedding (LDE), to manifold learning and pattern classification, in which the neighbor and class relations of data are used to construct the embedding for classification problems.
Proceedings ArticleDOI
Fusing generic objectness and visual saliency for salient object detection
TL;DR: Experimental results on two benchmark datasets demonstrate that the proposed model can simultaneously yield a saliency map of better quality and a more meaningful objectness output for salient object detection.
Journal ArticleDOI
Multiple Kernel Learning for Dimensionality Reduction
TL;DR: The proposed approach generalizes the framework of multiple kernel learning for dimensionality reduction, and distinguishes itself with the following three main contributions: first, the method provides the convenience of using diverse image descriptors to describe useful characteristics of various aspects about the underlying data, and consequently improves their effectiveness.
Proceedings ArticleDOI
From co-saliency to co-segmentation: An efficient and fully unsupervised energy minimization model
TL;DR: This work addresses two key issues of co-segmentation over multiple images by establishing an MRF optimization model that has an energy function with nice properties and can be shown to effectively resolve the two difficulties.
Proceedings ArticleDOI
Approximate tree matching and shape similarity
Tyng-Luh Liu,Davi Geiger +1 more
TL;DR: A framework for 2D shape contour (silhouette) comparison that can account for stretchings, occlusions and region information is presented and three local tree matching operations, merge, cut, and merge-and-cut, are introduced to yield optimally approximate matches.