Lifesize, Inc.

Lifesize Communications, Inc.

Stephen Flosdorf

Sulanka Goonetilleke

Jingqiang Li

Michael Vayden Jenkins

Patrick R McKinnon

Techniques for calibrating a video detection system are disclosed. A videoconferencing system may include a sensor configured to capture sequential frames of video image data during a videoconference. A processing subsystem may be configured to, using a multidimensional filter, generate data indicative of active areas of interest detected within respective ones of a plurality of video frames captured by the sensor, wherein the multidimensional filter is configured to identify active areas of interest (AAOIs) based at least on part upon a programmable density threshold. The processing subsystem may further: determine, based on the data indicative of AAOIs, that the videoconferencing system is in an unstable state; perform a calibration routine to identify one or more threshold values that reduce system instability; and apply the one or more threshold values to the multidimensional filter during further generation of data indicative of active areas of interest.

Dynamic calibration of detection system for active areas of interest within video data

Techniques for video processing using a topological map are disclosed. Program instructions stored on a computer-readable medium may be executable to receive data indicative of an active area of interest detected within a video frame captured by a detection system, and based upon one or more inclusion criteria, include the active area of interest within a stack of previously identified active areas of interest. The stack may occupy a primary frame within a two-dimensional video space. The instructions may be further executable to update a topological map of the two-dimensional video space based upon the stack, identify within the topological map a two-dimensional region that satisfies a density threshold, and adjust framing of a primary imaging camera based upon the two-dimensional region.

Video processing using topological map

Techniques for detecting an active area of interest within video data are disclosed. A videoconferencing system may include a sensor configured to capture sequential frames of video image data during a videoconference, and a processing subsystem. The processing subsystem may be configured to, for a given frame, apply filter(s) to individual pixels of the given frame to identify one or more pixels of interest within the given frame, where the filter(s) include a motion detection filter. The processing subsystem may further apply a multidimensional filter to the one or more pixels of interest, the multidimensional filter specifying one or more programmable density thresholds. The processing subsystem may further designate at least one group of pixels of interest that satisfies the one or more programmable density thresholds as an active area of interest; and control an aspect of the videoconference based at least in part upon the active area of interest.

Identifying active areas of interest within video data

A system and method for detecting local motion between image frames are presented. Each image frame may include a plurality of pixels. A first image frame and a second image frame may be received. Values of a first plurality of pixels at first pixel locations of the first image frame may be read. Values of a second plurality of pixels at second pixel locations of the second image frame may be read. The first pixel locations of the first image frame may correspond to the second pixel locations of the second image frame. A first delta function, based on differences between pixel values between the first and second image frames, may be calculated to produce a first output. The first output may be compared to a first one or more thresholds. It may be determined whether local motion has occurred in the second image frame based on said comparing.

Detection of local motion between image frames

A system and method for detecting defective pixels in a sensor. A plurality of pixel values of the sensor may be detected. The values may include those of a first pixel and each nearest neighboring pixel to the first pixel. A second pixel may have the highest value of the neighboring pixels. A third pixel may have the next highest value of the neighboring pixels. A first function may be performed on the second pixel value, producing a first output value. A second function may be performed on the third pixel value, producing a second output value. If the first pixel value is higher than the first output value, or, if the first pixel value is higher than the second output value and the second pixel value is higher than the second output value, it may be determined that the first pixel is defective.

Growing community of inventors

Round Rock, TX, United States of America

Stephen Flosdorf