The use of multilayer perceptrons (MLP) with threshold functions (binary step function activations) greatly reduces the complexity of the hardware implementation of neural networks...
Vassilis P. Plagianakos, George D. Magoulas, Micha...
To recover depth from images, the human visual system uses many monocular depth cues, which vision research has only begun to explore. Because a given image can have many possible...
Abstract—We experimentally evaluate bagging and seven other randomizationbased approaches to creating an ensemble of decision tree classifiers. Statistical tests were performed o...
Robert E. Banfield, Lawrence O. Hall, Kevin W. Bow...
Segmentation-based approach has shown significant success in stereo matching. By assuming pixels within one image segment belong to the same 3D surface, robust depth estimation ca...
We present a new syntactic parser that works left-to-right and top down, thus maintaining a fully-connected parse tree for a few alternative parse hypotheses. All of the commonly ...