Ricoh Company, Ltd.

Atsushi Shibata

Takashi Yano

Toshiharu Murai

Tatsuo Miyachi

Kazunori Hoshi

Osamu Adachi

Jun-ichiroh Fujimoto

The speaker recognition methods and systems of the current invention involves at least two processing units for performing the speaker recognition based upon his or her voice input. To perform the speaker recognition efficiently as well as securely, the voice input is initially processed at the input site so that intermediate voice characteristic information is extracted. The intermediate voice characteristic information is transmitted to a second location for the final determination for identifying or verifying a speaker.

Voice-based verification and identification methods and systems

A network interface unit couples a lattice communication network and at least one token ring network. The network interface unit includes a sequence controller which controls a token detector and a reception register so that message information received from the lattice communication network via a node apparatus connection port is looped back from the token ring network via a terminal connection port and output from the node apparatus connection port when an address detector detects that a destination address matches a stored source address, and absorbed when the address detector detects that the destination address does not match the stored source address. The network interface unit also includes a phantom signal part which outputs a detection signal via the terminal connection port when a phantom signal has been received, the phantom signal part including a set of resistors coupled to the terminal connection port, each of the resistors having a resistance that produces a detection signal having an appropriate level.

Network interface units and communication system using network interface

Growing community of inventors

Sagamihara, Japan

Atsushi Shibata