Mobile-cloud assisted video summarization framework for efficient management of remote sensing data generated by...

Area of Science:

• Medical Imaging
• Computer Science

Background:

• Wireless capsule endoscopy (WCE) generates large video datasets, requiring significant computational resources for analysis.
• Efficiently browsing and analyzing WCE data is crucial for remote health monitoring services.

Purpose of the Study:

• To develop a resource- and bandwidth-aware WCE video summarization framework.
• To reduce computational load and data transmission costs associated with WCE video analysis.

Main Methods:

• Implemented a framework for extracting representative keyframes by eliminating redundant and non-informative frames.
• Utilized Jeffrey-divergence for color histogram analysis and Boolean series for inter-frame correlation to remove redundancy.
• Employed multi-fractal texture features and an ensemble classifier to identify and remove non-informative frames.
• Incorporated a mobile-cloud architecture for adaptive offloading of summarization tasks to manage limited WCE resources.

Main Results:

• The proposed framework effectively extracts representative keyframes from WCE videos.
• Demonstrated significant savings in information transmission cost and bandwidth.
• Showcased a reduction in the time required for data analysts to browse WCE data.

Conclusions:

• The mobile-cloud based WCE video summarization framework offers an efficient solution for data analysis in remote health monitoring.
• The approach successfully addresses computational challenges and resource limitations of WCE systems.
• The framework enhances the usability of WCE by optimizing data browsing and reducing analytical overhead.