To leverage large-scale weakly-tagged images for computer vision tasks (such as object detection and scene recognition), a novel cross-modal tag cleansing and junk image filtering...
In this paper, we propose a new hybrid model for active contour image segmentation, which is able to segment non-uniform noisy images efficiently. The model is a combination betwe...
The main focus of image mining in the proposed method is concerned with the classification of brain tumor in the CT scan brain images. The major steps involved in the system are: p...
Abstract- We present a novel and computationally simple prediction stage in a Daubechies 5/3 ? like lifting structure for lossless image compression. In the 5/3 wavelet, the predic...
We present a framework for object detection that is invariant to object translation, scale, rotation, and to some degree, occlusion, achieving high detection rates, at 14 fps in c...
Michael Villamizar, Alberto Sanfeliu, Juan Andrade...