ABSTRACT In this paper, we report our experiments using a realworld image dataset to examine the effectiveness of Isomap, LLE and KPCA. The 1,897-image dataset we used consists of ...
Mei-Chen Yeh, I-Hsiang Lee, Gang Wu, Yi Wu, Edward...
The Random forest classifier comes to be the working horse for visual recognition community. It predicts the class label of an input data by aggregating the votes of multiple tree...
For many supervised learning tasks it is very costly to produce training data with class labels. Active learning acquires data incrementally, at each stage using the model learned...
We present a system for the estimation of unconstrained
3D human upper body movement from multiple cameras.
Its main novelty lies in the integration of three components:
single-...
Inverse multiplexing, or network striping, allows the construction of a high-bandwidth virtual channel from a collection of multiple low-bandwidth network channels. Striping syste...