Abstract: Techniques are presented for analyzing audio-video segments, usually from multiple sources. A combined similarity measure is determined from text similarities and video similarities. The text and video similarities measure similarity between audio-video scenes for text and video, respectively. The combined similarity measure is then used to determine similar scenes in the audio-video segments. When the audio-video segments are from multiple audio-video sources, the similar scenes are common scenes in the audio-video segments. Similarities may be converted to or measured by distance. Distance matrices may be determined by using the similarity matrices. The text and video distance matrices are normalized before the combined similarity matrix is determined. Clustering is performed using distance values determined from the combined similarity matrix.

Abstract: A voice-mail annotation system and method that enables callers to annotate the messages they leave at the time they leave them. Annotation may be performed using a touch screen or a telephone key-pad. In addition to allowing the caller to more clearly specify the intent of their message and the location of crucial information within the message, the annotations of this invention then allow the person retrieving the message flexibility in how they listen to the message. These choices include prioritizing the messages by the importance of subject, listening to summaries of the message and having random access to embedded information such as names and call back numbers.

Abstract: A method and apparatus for interactive language instruction is provided that displays text files for processing, provide key features and functions for interactive learning, displays facial animation, and provides a workspace for language building functions. The system includes a stored set of language rules as part of the text-to-speech sub-system, as well as another stored set of rules as applied to the process of learning a language. The method implemented by the system includes digitally converting text to audible speech, providing the audible speech to a user or student (with the aid of an animated image in selected circumstances), prompting the student to replicate the audible speech, comparing the student's replication with the audible speech provided by the system, and providing feedback and reinforcement to the student by, for example, selectively recording or playing back the audible speech and the student's replication.

Abstract: A voice-mail annotation system and method that enables callers to annotate the messages they leave at the time they leave them. Annotation may be performed using a touch screen or a telephone key-pad. In addition to allowing the caller to more clearly specify the intent of their message and the location of crucial information within the message, the annotations of this invention then allow the person retrieving the message flexibility in how they listen to the message.

Abstract: ? Techniques are presented for analyzing audio-video segments, usually from multiple sources. A combined similarity measure is determined from text similarities and video similarities. The text and video similarities measure similarity between audio-video scenes for text and video, respectively. The combined similarity measure is then used to determine similar scenes in the audio-video segments. When the audio-video segments are from multiple audio-video sources, the similar scenes are common scenes in the audio-video segments. Similarities may be converted to or measured by distance. Distance matrices may be determined by using the similarity matrices. The text and video distance matrices are normalized before the combined similarity matrix is determined. Clustering is performed using distance values determined from the combined similarity matrix.

Abstract: A method and apparatus for automatically controlling the operation of a speech recognition system without requiring unusual or unnatural activity of the speaker by passively determining if received sound is speech of the user before activating the speech recognition system. A video camera and microphone are located in a hand-held device. The video camera records a video image of the speaker's face, i.e., of speech articulators of the user such as the lips and/or mouth. The recorded characteristics of the articulators are analyzed to identify the sound that the articulators would be expected to make, as in “lip reading”. A microphone concurrently records the acoustic properties of received sound proximate the user. The recorded acoustic properties of the received sound are then compared to the characteristics of speech that would be expected to be generated by the recorded speech articulators to determine whether they match.

Abstract: A method and apparatus for interactive language instruction is provided that displays text files for processing, provide key features and functions for interactive learning, displays facial animation, and provides a workspace for language building functions. The system includes a stored set of language rules as part of the text-to-speech sub-system, as well as another stored set of rules as applied to the process of learning a language. The method implemented by the system includes digitally converting text to audible speech, providing the audible speech to a user or student (with the aid of an animated image in selected circumstances), prompting the student to replicate the audible speech, comparing the student's replication with the audible speech provided by the system, and providing feedback and reinforcement to the student by, for example, selectively recording or playing back the audible speech and the student's replication.

Abstract: A method for simulating natural movement of a computer-generated synthesized talking head using an N-state Markov model. Initially, the number of states N is set representing N different positional orientations of the synthesized talking head. The transitions from each possible current position to each possible next position of the talking head are identified with a set of transitional probabilities in a range from zero to one. A uniformly distributed random number is generated and the state identified by the transition probabilities associated with the random number identifies the next position of the talking head. The synthesized talking head is rotated to that next position and remains there for a durational time period before transitioning to a next position which is probabilistically determined in like manner.