Advancing content-based image retrieval by exploiting image color and region features

Silence detection and removal is an essential building block of any multimedia video conferencing system. It reduces the bandwidth requirements of the underlying network transport service and helps to maintain an acceptable end-to-end delay for audio. We analyze the requirements for a silence detection algorithm hosted on a multimedia communication system, and propose a novel low-complexity algorithm operating in the non-linear (mo)-law domain. After discussing the constraints which are imposed by the architecture of the system hardware (computer, packet-based network), we show that several recently proposed silence detection algorithms fail to meet all of these constraints. A new approach is then introduced, based on the small- and large-signal behavior of the speech waveform in the (mo)-law domain. The new algorithm is compared with a recent design that meets several of our requirements; experimental results indicate that it performs significantly better in the particular environment at hand.