Cloud Video Synopsis Generation with Object Indicators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for generating synopsis videos from network cameras are expensive, require local processing servers, and make it difficult for users to navigate and select objects of interest due to large video streams and overlapping objects.
Innovation Solution
A system that records video files from network cameras and processes them in a cloud server, generating synopsis videos with chapters and object indicators, allowing users to efficiently navigate and select objects of interest without the need for local processing equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If local processing servers are used to generate synopsis videos from large recorded video streams, then synopsis videos can be generated, but the system becomes expensive and difficult to install
Solution Approach 1:
The patent introduces cloud-based processing servers as an intermediary between the network cameras and user devices. The cameras transmit video streams to the cloud servers, which perform the complex synopsis generation tasks remotely. This mediator approach allows consumers to access advanced video analysis capabilities without needing to purchase or maintain expensive local processing equipment, thereby resolving the contradiction between functionality and system complexity.
2Reliability
If conventional systems process large video streams locally, then synopsis videos can be generated, but it is difficult for users to sort through and select objects of interest
Solution Approach 1:
The patent segments the continuous video stream into discrete object-based units. The cloud processing servers detect and track individual objects throughout the video, creating separate object records with timestamps, locations, and visual data. This segmentation transforms the overwhelming continuous stream into manageable, individually addressable objects that users can easily browse and select through graphical interfaces, resolving the contradiction between generating comprehensive synopses and enabling easy object selection.
Solution Approach 2:
The patent adds a new dimension to video navigation by creating object-centric indexing and presentation layers. Instead of traditional temporal or spatial navigation, users can search and filter objects by multiple dimensions including object type, detected actions, time ranges, and locations. This multi-dimensional organization system, implemented through cloud processing and user interface displays, makes it significantly easier to locate and select objects of interest within large video datasets.
3Quantity of substance
If cloud servers process video files, then data transfer and storage requirements are reduced for consumers, but processing must be done remotely
Solution Approach 1:
The patent employs cloud servers as remote intermediaries that handle all heavy processing tasks including video analysis, object detection, and synopsis generation. Consumers only need to upload their video files to the cloud and receive processed results, avoiding the need to download and process terabytes of raw video data locally. This intermediary approach minimizes data transfer volumes while maintaining full processing capability through remote cloud infrastructure.
Data Source
AI summary
Systems and methods are disclosed for efficient generation and presentation of synopsis videos that include video images of multiple objects captured at various times displayed at a common time. The system may include one or more network cameras that, when a moving object is detected, record a video file for that object and provide the video file to a cloud server. The cloud server may queue video files from the one or more network cameras and, when a user requests a synopsis video of a period of time, the video files captured during that period of time may be used by the cloud server to generate the synopsis video. The cloud server may also generate a plurality of object indicators that are provided with each of several chapters of the synopsis video in a scrollable arrangement that allows the user the scroll through objects and chapters.