Method and system for foreground detection using multi-modality fusion graph cut
AssigneeInstitute for Information Industry
InventorsFelix L. Chang, Winston H. Hsu, Miao-Fen Chueh
Life of this patent
Estimating the term…
Abstract
A method for foreground detection using multi-modality fusion graph cut includes the following steps. A video frame of a video sequence is inputted. A foreground region in the video frame is designated using adaptive background Gaussian Mixture Model and a threshold value. The foreground region is segmented using a multi-modality fusion graph cut procedure. A computer program product using the method and a system for foreground detection using multi-modality fusion graph cut are also disclosed herein.
Claims18
A method for foreground detection using multi-modality fusion graph cut comprising the following steps:
- inputting a video frame of a video sequence;
- designating a foreground region in the video frame using an adaptive background Gaussian Mixture Model and a threshold value, wherein the foreground region comprises at least one object block, furthest comprising:
- saving the information regarding each object block using the adaptive background Gaussian Mixture Model;
- comparing each object block with the adaptive background Gaussian Mixture Model;
- and
- updating the adaptive background Gaussian Mixture Model or generating another adaptive background Gaussian Mixture Model according to the trend of a change when the adaptive background Gaussian Mixture Model detects the change;
- and
- segmenting the foreground region using a Multi-Modality Fusion Graph Cut procedure.
A system for foreground detection using multi-modality fusion graph cut comprising:
- an input means for inputting a video frame of a video sequence;
- a foreground designation means for designating a foreground region in the video frame using an adaptive background Gaussian Mixture Model and a threshold value, wherein the foreground region comprises at least one object block, further comprising:
- a saving means for saving the information regarding each object block using the adaptive background Gaussian Mixture Model;
- a comparing means for comparing each object block with the adaptive background Gaussian Mixture Model;
- and
- a model update means for updating the adaptive background Gaussian Mixture Model or generating another adaptive background Gaussian Mixture Model according to the trend of a change when the adaptive background Gaussian Mixture Model detects the change;
- and
- a foreground segmentation means for segmenting the foreground region using a Multi-Modality Fusion Graph Cut procedure.
A non-transitory computer-readable medium comprising program codes for executing the following steps:
- inputting a video frame of a video sequence;
- designating a foreground region in the video frame using an adaptive background Gaussian Mixture Model and a threshold value, wherein the foreground region comprises at least one object block, further comprising:
- saving the information regarding each object block using the adaptive background Gaussian Mixture Model;
- comparing each object block with the adaptive background Gaussian Mixture Model;
- and
- updating the adaptive background Gaussian Mixture Model or generating another adaptive background Gaussian Mixture Model according to the trend of a change when the adaptive background Gaussian Mixture Model detects the change;
- and
- segmenting the foreground region using a Multi-Modality Fusion Graph Cut procedure.
Dependencies are inferred from the claim text.