Video Summarization via Sparse Basis Function Combination

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video summarization methods rely on key frame extraction, which is susceptible to noise and non-linearity, and are not data adaptive, limiting their effectiveness across various video content types.

Innovation Solution

A method that defines a global feature vector for a video sequence, selects subsets of frames, extracts frame feature vectors, determines a sparse combination of basis functions, and forms a video summary without relying on key frames, incorporating low-level image quality and high-level semantic information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If key frame extraction algorithms are used for video summarization, then the summarization process can be simplified, but the performance becomes susceptible to noise and non-linearity

Engineering Contradiction:

Improvesimplicity of summarization processVSAvoidperformance reliability

Core Design Contradiction:

Ease of manufactureVSReliability

Solution Approach 1:

The patent replaces traditional mechanical key frame extraction algorithms with a sparse representation model based on signal processing theory. Instead of using heuristic-based frame selection, the invention formulates video summarization as a sparse coding problem where the global feature vector is represented as a sparse linear combination of basis function vectors extracted from video frames, thereby eliminating susceptibility to noise and non-linearity while maintaining process simplicity

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The invention changes the fundamental parameters of video summarization by transitioning from discrete key frame selection to continuous sparse coefficient optimization. The method transforms the problem into finding optimal sparse coefficients that minimize reconstruction error between the global feature vector and the linear combination of basis functions, thereby improving reliability through mathematically grounded parameter optimization

Inventive Principle:

Principle #35Parameter changes

2Device complexity

If traditional key frame-based methods are used, then the computational process is simpler, but the method is not adaptive to different content types

Engineering Contradiction:

Improvecomputational complexityVSAvoidadaptability to different content types

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent introduces dynamics into the summarization process by making the basis function selection adaptive to different video content types. The method dynamically adjusts which video frames serve as basis functions based on the specific characteristics of each video, allowing the system to adapt to various content types while maintaining a unified computational framework through sparse representation

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The invention creates a universal video summarization framework that can handle different content types through the sparse representation model. The same mathematical formulation applies to all video types, but the basis functions are selected adaptively from the specific video content, providing both universality in approach and adaptability in application across diverse video genres

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Measurement precision

If more video frames are selected as key frames, then the summary quality improves, but the temporal redundancy increases

Engineering Contradiction:

Improvesummary qualityVSAvoidtemporal redundancy

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the essential information needed for video summarization by formulating the problem as sparse representation. Instead of selecting multiple key frames, the method extracts a sparse set of basis function vectors and computes minimal coefficients to reconstruct the global feature vector, thereby capturing summary quality while minimizing temporal redundancy through selective extraction of essential components

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS8467610B2Video summarization using sparse basis function combination

Publication Date: 2013.06.18 KODAK ALARIS LLC

AI summary

A method for determining a video summary from a video sequence including a time sequence of video frames, comprising: defining a global feature vector representing the entire video sequence; selecting a plurality of subsets of the video frames; extracting a frame feature vector for each video frame in the selected subsets of video frames; defining a set of basis functions, wherein each basis function is associated with the frame feature vectors for the video frames in a particular subset of video frames; using a data processor to automatically determine a sparse combination of the basis functions representing the global feature vector; determining a summary set of video frames responsive to the sparse combination of the basis functions; and forming the video summary responsive to the summary set of video frames.