Abstract: Multi-label active learning may entail training a classifier with a set of training samples having multiple labels per sample. In an example embodiment, a method includes accepting a set of training samples, with the set of training samples having multiple respective samples that are each respectively associated with multiple labels. The set of training samples is analyzed to select a sample-label pair responsive to at least one error parameter. The selected sample-label pair is then submitted to an oracle for labeling.

Abstract: An automated system and method for broadcasting meetings over a computer network. The meeting is filmed using an omni-directional camera system and capable of being presented to a viewer both live and on-demand. The system of the present invention includes an automated camera management system for controlling the camera system and an analysis module determining the location of meeting participants in the meeting environments. The method of the present invention includes using the system of the present invention to broadcast an event to a viewer over a computer network. In particular, the method includes filming the event using an omni-directional camera system. Next, the method determines the location of each event participant in the event environment. Finally, a viewer is provided with a user interface for viewing the broadcast event. This user interface allows a viewer to choose which event participant that the viewer would like to view.

Abstract: A “virtual video studio”, as described herein, provides a highly portable real-time capability to automatically capture, record, and edit a plurality of video streams of a presentation, such as, for example, a speech, lecture, seminar, classroom instruction, talk-show, teleconference, etc., along with any accompanying exhibits, such as a corresponding slide presentation, using a suite of one or more unmanned cameras controlled by a set of videography rules. The resulting video output may then either be stored for later use, or broadcast in real-time to a remote audience. This real-time capability is achieved by using an abstraction of “virtual cameramen” and physical cameras in combination with a scriptable interface to the aforementioned videography rules for capturing and editing the recorded video to create a composite video of the presentation in real-time under the control of a “virtual director.

Abstract: Correlative multi-label image annotation may entail annotating an image by indicating respective labels for respective concepts. In an example embodiment, a classifier is to annotate an image by implementing a labeling function that maps an input feature space and a label space to a combination feature vector. The combination feature vector models both features of individual ones of the concepts and correlations among the concepts.

Abstract: A program distribution system includes a plurality of set-top boxes that receive broadcast programming and segmentation data from content and information providers. The segmentation information indicates portions of programs that are to be included in skimmed or condensed versions of the received programming, and is produced using manual or automated methods. Automated methods include the use of ancillary production data to detect the most important parts of a program. A user interface allows a user to control time scale modification and skimming during playback, and also allows the user to easily browse to different points within the current program.

Abstract: A system and method for teleconferencing and recording of meetings. The system uses a variety of capture devices (a novel 360° camera, a whiteboard camera, a presenter view camera, a remote view camera, and a microphone array) to provide a rich experience for people who want to participate in a meeting from a distance. The system is also combined with speaker clustering, spatial indexing, and time compression to provide a rich experience for people who miss a meeting and want to watch it afterward.

Abstract: An improved image retrieval process based on relevance feedback uses a hierarchical (per-feature) approach in comparing images. Multiple query vectors are generated for an initial image by extracting multiple low-level features from the initial image. When determining how closely a particular image in an image collection matches the initial image, a distance is calculated between the query vectors and corresponding low-level feature vectors extracted from the particular image. Once these individual distances are calculated, they are combined to generate an overall distance that represents how closely the two images match. According to other aspects, relevancy feedback received regarding previously retrieved images is used during the query vector generation and the distance determination to influence which images are subsequently retrieved.

Abstract: Automatic detection and tracking of multiple individuals includes receiving a frame of video and/or audio content and identifying a candidate area for a new face region in the frame. One or more hierarchical verification levels are used to verify whether a human face is in the candidate area, and an indication made that the candidate area includes a face if the one or more hierarchical verification levels verify that a human face is in the candidate area. A plurality of audio and/or video cues are used to track each verified face in the video content from frame to frame.

Abstract: A system and method for teleconferencing and recording of meetings. The system uses a variety of capture devices (a novel 360° camera, a whiteboard camera, a presenter view camera, a remote view camera, and a microphone array) to provide a rich experience for people who want to participate in a meeting from a distance. The system is also combined with speaker clustering, spatial indexing, and time compression to provide a rich experience for people who miss a meeting and want to watch it afterward.

Abstract: Automatic detection and tracking of multiple individuals includes receiving a frame of video and/or audio content and identifying a candidate area for a new face region in the frame. One or more hierarchical verification levels are used to verify whether a human face is in the candidate area, and an indication made that the candidate area includes a face if the one or more hierarchical verification levels verify that a human face is in the candidate area. A plurality of audio and/or video cues are used to track each verified face in the video content from frame to frame.

Abstract: A decentralized computer network architecture and method that gathers metadata from local and remote clients and, based on that metadata, locally makes a decision whether to send a packet over the network. Each client listens to what other clients are doing, and only sends when the total number of concurrent speakers is below some threshold. In a multi-party voice conferencing embodiment, the threshold is a number of concurrent speakers that is restricted to less than a certain number. Under the decentralized computer network architecture, the type of network topology used to connect the clients is flexible, as long as each client is running a peer-aware system to decide locally whether to send their packets. The decentralized computer network architecture and method is distributed to run on each client, making it suitable for a wide variety of network topologies (such as full-mesh, bridge-based, or a hybrid of the two).

Abstract: Audio/video programming content is made available to a receiver from a content provider, and meta data is made available to the receiver from a meta data provider. The meta data corresponds to the programming content, and identifies, for each of multiple portions of the programming content, an indicator of a likelihood that the portion is an exciting portion of the content. In one implementation, the meta data includes probabilities that segments of a baseball program are exciting, and is generated by analyzing the audio data of the baseball program for both excited speech and baseball hits. The meta data can then be used to generate a summary for the baseball program.

Abstract: An event-based system and process for recording and playback of collaborative electronic presentations is presented. The present system and process includes a technique for recording collaborative electronic presentations by capturing and storing the interactions between each participant and presentation data where each interaction event is timestamped and linked to a data file comprising the presentation data. The present system and process also includes a technique for playing back the recorded collaborative electronic presentation, which involves displaying the presentation data in an order it was originally presented and reproducing the recorded interactions between each participant and the displayed presentation data at the same point in the presentation that they were originally performed, based on the aforementioned timestamps.

Abstract: An automated camera management system and method for capturing presentations using videography rules. The system and method use technology components and aesthetic components represented by the videography rules to capture a presentation. In general, the automated camera management method captures a presentation using videography rules to determine camera positioning, camera movement, and switching or transition between cameras. The videography rules depend on the type of presentation room and the number of audio-visual camera units used to capture the presentation. The automated camera management system of the invention uses the above method to capture a presentation in a presentation room. The system includes a least one audio-visual (A-V) camera unit for capturing and tracking a subject based on vision or sound. The (A-V) camera unit includes any combination of the following components: (1) a pan-tilt-zoom (PTZ) camera; (2) a fixed camera; and (3) a microphone array.

Abstract: An automated system and method for producing videos using expert video production rules. The automated video production system and method of the present invention is particularly well-suited for the online publishing of lectures. The system of the present invention includes a camera system for capturing the lecture, such as an audience-tracking camera and a non-intrusive lecturer-tracking camera. Moreover, each camera may be controlled by a cinematographer. Tracking of a subject is performed using a history-based, reduced-motion tracker that sets up a camera shot based on the subject's movement history and leaves the shot fixed until the camera is switched. The system also includes a virtual director module having a probabilistic finite state machine (FSM) module that uses probabilistic rules and the expert video production rules to determine a current camera view.

Abstract: A system and process for providing a network and computer-based integrated development environment is presented that provides collaboration and information sharing for development project team members. Generally, this is accomplished by integrating a presence and activity awareness information module, a collaboration tools module, and a user interface into a single environment that is accessible over a distributed network and serves as a virtual development complex. The information module continuously collects, monitors and analyzes information about the presence of each team member and their activity in the project. The tools module provides a wide range of facilities for synchronous and asynchronous collaboration and information sharing between team members.

Abstract: Systems and methods for detecting people or speakers in an automated fashion are disclosed. A pool of features including more than one type of input (like audio input and video input) may be identified and used with a learning algorithm to generate a classifier that identifies people or speakers. The resulting classifier may be evaluated to detect people or speakers.

Abstract: A program distribution system includes a plurality of set-top boxes that receive broadcast programming and segmentation data from content and information providers. The segmentation information indicates portions of programs that are to be included in skimmed or condensed versions of the received programming, and is produced using manual or automated methods. Automated methods include the use of ancillary production data to detect the most important parts of a program. A user interface allows a user to control time scale modification and skimming during playback, and also allows the user to easily browse to different points within the current program.

Abstract: Background blurring is an effective way to both preserve privacy and keep communication effective during video conferencing. The present image background blurring technique is a light weight real-time technique to perform background blurring using a fast background modeling procedure combined with an object (e.g., face) detector/tracker. A soft decision is made at each pixel whether it belongs to the foreground or the background based on multiple vision features. The classification results are mapped to a per-pixel blurring radius image to blur the background. In another embodiment, the image background blurring technique blurs the background of the image without using the object detector.

Abstract: A system and process is described for estimating the location of a speaker using signals output by a microphone array characterized by multiple pairs of audio sensors. The location of a speaker is estimated by first determining whether the signal data contains human speech components and filtering out noise attributable to stationary sources. The location of the person speaking is then estimated using a time-delay-of-arrival based SSL technique on those parts of the data determined to contain human speech components. A consensus location for the speaker is computed from the individual location estimates associated with each pair of microphone array audio sensors taking into consideration the uncertainty of each estimate. A final consensus location is also computed from the individual consensus locations computed over a prescribed number of sampling periods using a temporal filtering technique.