Flashcards
Video Analysis and Machine Learning
Cards in this deck (15)
- 11. Explain when and how machine learning and computer vision is useful in industry, public institutions and research1. Explain when and how machine learning and computer vision is useful in industry, public institutions and research
- 2What is Video Surveillance?Process of using video cameras to monitor behaviour/activities of people used to protection, influence and direct
- 3What does Surveillance mean?- To monitor and observe individuals movements, conversations and other activities or communication
- Recording suspicious behavior - 4Why Video Surveilance?-It ensures community safety
-Secures evidence for courts
-Gathers intelligence on criminals/terrorist activity
Helps to apprehend increasingly sophisticated offenders
Even if cameras aren't working/real they can influence a person to act differently i.e. not commit a crime as the camera will catch them - 5What are the Different Categories of collecting Video Surveillance?Recording: is used to collect info for evidence
Passive: an employee may monitor screens on the side of another task - they swap - the human does the work
Active: may include using AI - automatically monitoring an area for assisting security officers - 6What does the Analyst Do? And apply it to the Image Analyst and Video Analyst Workstations.Uses image enhancement software/hardware
Makes extraction of information from the image easier
They require less skills than now due to advance technology
Image Analyst Workstation
Extracts information from visual data i.e. a frame or video
Must be highly Qualified
Video Surveillance workstation
Must be a Mediumly qualified analyst
Watches multiple video screens at once
Real time viewing - 7What are the Cyber Physical Security Risks?Cyber security is the measurements taken to protect computer assets
Includes things like phishing attack and inside intruders where the person gets onto the computer in person this may mean using biometrics to make it safer - 8Name 3 different uses of Video AnalyticsDefence Reconnaisance
Medical Diagnosis
Fall Detector
Sports Analysis
VR Games
Scientific Data processing - 9What are the different areas of defence reconnaissance?Target Detection
Where as humans lose focus after 15 minutes, this does not happen when the surveillance is automated, this is particularly important for missions as they are 4 hours
It can then cue the Analyst when it has detected something and they make the final decision
Target Tracking
If a target is detected it decides if it's a friend or foes and will track the moving target but the analyst needs to keep moving target in the centre of the sensor but this is difficult to manoeuvre into place.
Surveillance/Reconnaissance
Automated Tracking, analyst simply clicks on the screen at the target position - 10How is Video Analytics used in the medical field?Can also be used for medical diagnosis like detecting for cancer in order to define if a person has cancer from the slides from the biopsy
The biopsy is analysed under a microscope by a pathologist they look at the textures and structures - 11How can it be used in Nursing Homes?Can be used for fall detector for nursing homes or detecting if resident leaves
- 12Why and how would you use Automated proccessing rather than Human?Sometimes it must be done in real time and cannot wait: Video surveillance i.e. catching criminals ect
Often the analysis of images is boring and tiring
Analysts are expensive to hire
Allowing automated processing can assist an analyst i.e pre screening medical images or triggering an alarm for a security officer but in the end the analyst makes the final decision. - 13What is Computer Vision?Computer vision is used for things like translating signs, face filters, cameras tracking faces and focusing etc. There are things like automated driving potentially coming soon
But its difficult for computers as whereas we see colurs and how the colours go together they see numbers
In some cases pre processing the image may help as it can enhance the image, normalize all images can do this via. Contrast Enhancement, Equalization, Noise Reduction - 14In the context of VAML what is the difference between Data and Information?The data is the raw unprocessed visual data i.e. the photo of a biopsy slide
The information is the contextualized image with meaningful output i.e. defining that the biopsy slide has cancer
Or Given the Data - fall detector the information may be Standing or Fallen - 15What are the challenges of Automated Processing?Not good cameras, too many people in the frame, doing it in real time, but it can also help out a lot.
Information
- Description
- University Video Analysis and Machiene Learning Flashcards on Video Analysis and Machine Learning, created by Abbie Welch on 28/09/2025.
- Created by
- Abbie Welch
- Views
- 1
- Rating