The use of visual information derived from accurate lip extraction, can provide features invariant to noise perturbation for speech recognition systems and can be also used in a w...
We give a fast rejection scheme that is based on image segments and demonstrate it on the canonical example of face detection. However, instead of focusing on the detection step w...
FastICA is arguably one of the most widespread methods for independent component analysis. We focus on its deflation-based implementation, where the independent components are ext...
Extraction and utilization of high-level semantic features are critical for more effective video retrieval. However, the performance of video retrieval hasn't benefited much d...
A method for identifying JPEG 2000 images with different coding parameters, such as DWT-filters, code-block sizes, quantization step sizes, and resolution levels, is presented. ...