Browse Prior Art Database

Finding a Fight Scene in a Police Video

IP.com Disclosure Number: IPCOM000247613D
Publication Date: 2016-Sep-20
Document File: 1 page(s) / 20K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to use analytics and cognitive APIs and stream processing running on a cloud platform to accelerate the process of, and ensure accuracy for action in emergency scenarios including: elderly patient care, law enforcement video review, natrual and human-induced disasters etc

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 01 of 1

Finding a Fight Scene in a Police Video

With the advent a wider array of camera networks (CCTV, police body cameras, smartph cameras), ideally, the camera is always on for the for the field of view in each of the above scenarios. Such footage combined with cognitive API triggers and spark/MR-based analytics engines facilitate detection of anomalous events as applied to elderly patient care, city/town/municipal management and law enforcement to mention a few.

The exiting process flow is:


1. Camera


2. Video data ingestion via streaming engine

3. Cloud-based cognitive API triggers to detect an anomalous event if one occurs (prior system training for such data would have had to occur)

4. Pertinent cloud-based storage (cloudant)

5. Post priori spark/MR-analytics-based data processing to refine original conclusions of anomalous event
6. Human review

The novel contribution is a method to insert cognitive processing to detect anomalous events.

The cognitive analysis system can:

1. Perform image recognition and matching to identify segments in which a person-to-person interaction takes place

A. Redact footage with no person-to-person interaction


B. Forward footage with person-to-person interaction

2. Perform emotion analysis and tone organizing on the forwarded video to identify

whether the video shows a possible incident involving an escalation or conflict

1