System and method for adaptive video fast forward using scene generative models
First Claim
1. A system for automatically identifying image frames in an image sequence, comprising:
- providing a first image sequence of at least one image frame;
automatically training a generative model on the first image sequence;
providing a second image sequence of at least one image frame; and
automatically determining a likelihood of at least one frame of the second image sequence under the generative model.
2 Assignments
0 Petitions
Accused Products
Abstract
Computationally efficient searching, browsing and retrieval of one or more objects in a video sequence are accomplished using learned generative models. The generative model is trained on an automatically or manually selected query sequence from a sequence of image frames. The resulting generative model is then used in searching, browsing or retrieval of one or more similar or dissimilar image frames or sequences within the image sequence by determining the likelihood of each frame under the learned generative model. Further, this method allows for automatic separation and balancing of various causes of variability while analyzing the image sequence. The generative models are based on appearances of multiple, possibly occluding objects in an image sequence. Further, the search strategies used include clustering and intelligent fast forward through the image sequence. Additionally, in one embodiment, a fast forward speed is relative to the current frame likelihood under the learned generative model.
46 Citations
49 Claims
-
1. A system for automatically identifying image frames in an image sequence, comprising:
-
providing a first image sequence of at least one image frame; automatically training a generative model on the first image sequence; providing a second image sequence of at least one image frame; and automatically determining a likelihood of at least one frame of the second image sequence under the generative model. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)
-
-
26. A computer-implemented process for automatically identifying similar image frames in one or more image sequences, comprising:
-
acquire at least one image sequence, each image sequence having at least one image frame; select a query sample consisting of at least one image frame from one of the at least one image sequences; input a desired number of blobs to be modeled in generative model of the query sample; automatically learn a generative model from the query sample, wherein the generative model includes a set of model parameters that represent the desired number of blobs; and compare the frames in each image sequence to the generative model to determine a likelihood of each frame under the generative model. - View Dependent Claims (27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38)
-
-
39. A computer-readable medium having computer executable instructions for automatically determining a playback speed for an image sequence, comprising:
-
selecting a query sample consisting at least one image frame; automatically learning a generative model from the query sample; providing at least one image sequence for playback; comparing the frames in the at least one image sequence provided for playback to the generative model to determine a likelihood of each frame under the generative model; and automatically varying a playback speed of the at least one image sequence in inverse proportion to the likelihood of each frame. - View Dependent Claims (40, 41, 42, 43, 44, 45, 46, 47, 48, 49)
-
Specification