This site uses cookies to improve customer service and for other purposes.
For more information, please click here.

Technology

  1. HOME
  2. Technology
  3. Video Summarization

Video Summarization

Morpho Video Summary™

“MorphoVideo Summary™”(*1) is a video summarization technology that uses AI to extract only the appropriate scenes from vast amounts of video data to meet your requirements, thereby reducing data volume.

 

*Please refer to this press release regarding the collaboration with Micware’s Mvcube.

Use Case

  • Self-driving / On-board camera
  • Video production
  • Ship / Train / Airplane

Function or Purpose

  • Object detection

Features

  • Various video analysis engines along the customer application
  • Reduce data management costs by reducing data volume
  • Intuitive operation by dedicated GUI

Product Details

Development Background

Video from dash cams and other in-vehicle cameras is used not only as evidence of accidents, but also for a wide range of purposes; from road inspections to technological development of advanced driver assistance/automatic driving technology. However, the sheer volume of data imposes high maintenance cost and verification man-hours for automobile manufactures and related companies. As a solution, Morpho has developed “Morpho Video Summary,” which uses AI to summarize video data.

Features

Depending on the customer application, the requirements for summarizing video vary widely. In road inspection applications, the customer would want to keep high priority on scenes where cracks or other defects in the road can be seen. In the development of automated driving technology, you would want to collect mainly scenes that show vehicles or people. Thus, the required summarization rules are quite different depending on the final video application.

 

To cover diverse requirements, “Morpho Video Summary” provides a variety of video analysis engines for users to set up detailed summary rules according to their applications. For example, the “Driving Scene Detection Engine” can determine scenes in the video where the car is in motion. This allows users to create a summary rule that eliminates scenes in the video where their car is stopped at a red light (scenes with little change on the screen). In addition, video analysis results from various other detection engines can be combined and prioritized by each engine to realize complex summary rule settings tailored to the application.

Typical engines available

– Driving scene detection

– Vehicle, pedestrian, traffic signal, and motorcycle scene detection

– Road deterioration scene detection

*Additional engines will be available soon.

 

“Morpho Video Summary” provides a dedicated GUI (*2) for users to create these summary rules with visual aid. These summary rules can be saved and then applied to all videos in a batch process(*3).

 

Demonstration of video summarization using “Morpho Video Summary”

“Morpho Video Summary” has been added to the standard AI technologies in the “Morpho Automotive Suite™,” an AI solution package for the automotive field launched in May 2023. This new product will contribute to solving issues in the automotive field and adding value to customers’ products by providing not only stand-alone AI technology but also additional AI learning tailored for customers and proposing product customization for each application.

 

*1: The product name “Morpho Video Summary” may change in the future.

*2: GUI is “Graphical User Interface. It is a screen that can be operated intuitively with a mouse or other pointing device using graphics such as icons and buttons.

*3: Batch processing is to collect a certain amount of data for a certain period of time and perform batch processing.

The Format to Provide the Product

Morpho Video Summary" runs on the customer's on-premise server or on a cloud server provided by Morpho. Communication with the server is via webAPI, which users access from their client PCs.

Therefore, customers can develop their own GUI using the webAPI.

format