Click me

Human Eye vs Vision Systems

FACE OFF VS. THE HUMAN EYE COMPUTER VISION SYSTEMS Computer vision and machine vision systems have made huge technological leaps and bounds in the past decade. Used in everything from traffic cameras to quality inspection in manufacturing to medical im aging, vision systems lend a precision and efficiency to applications that humans find tough to beat. So when the human eye and a vision system go toe-to-toe in several scenarios, which would emerge victorious? ELEMENTS OF IMAGE PR OCESSING SYSTEMS VS. THE HUMAN EYE HUMAN EYE FUNCTION COMPUTER The pupil and iris controls light entry into the eye. The sensor intercepts and controls the radiant energy of IMAGE SENSOR the object. The retina focuses and projects the image onto the back of the eye. The digitizer produces images from discrete intensity IMAGE TRANSFER values. The cones (for color) and the rods Using an algorithm, a computer processes the (for low-light contrast) convert the image into neural signals for the brain to translate the images. IMAGE PROCESSER image. Has short-term (with frame buffers) and Archival Storage (optical/magnetic disks). The front of the brain manages short-term memory, the back of the STORAGE brain processes visual info. The brain "displays" images for us, Images displayed through various outlet such as TVs, CRTS, etc. but must communicate through DISPLAY/RECORDER words or illustrations for others to view. VISION & IMAGING SYSTEMS AT W ORK 1 HIGH-END SURVEILLANCE Gone are the days of grainy VHS security footage and unreliable eyewitness testimony. With high res CCD/CMOS cameras, police have a much easier time finding reliable "witnesses" to collect evidence for crimes. 03 SURVEILLANCE CAMERAS DON'T JUST RECORD WHAT THEY SEE; THEY'RE PROCESSING IT AND TURNING IT INTO DOCUMENTED DATA, SUCH AS: 02 FACIAL RECOGNITION Based off of facial nodal points (ie: shape of nose), which help pair footage with police databases. OBJECT BASED SEARCHES SCRNING PROGRESS 87% These allow operators to search through weeks of videos in seconds, looking for specific objects. LICENSE PLATE READERS (LPRS) Police officers average a few dozen per shift, LPRS average 10,000. OVER 3,000+ IN 2008, CCTV FOOTAGE HELPED SOLVE CCTV CAMERAS 65 86 MONITOR LOWER AND MIDTOWN MANHATTAN REAL LIFE: The 2013 Boston Marathon bombing investigation took a quick turn from officials having "a few leads" to "possible arrests" after pulling surveillance footage from nearby businesses that caught suspects on camera at various times throughout the run. OUT OF In similarly chaotic situations, eyewitness testimonies often conflict with one another. Our vision is subject to the interpretation of our brains, and different people may notice or perceive completely different things. MURDER INVESTIGATIONS IN THE UK WINNER: VISION SYSTEM 2 TRAFFIC CAMERAS You don't need a cop around to get caught running a red light. As traffic cameras become more commonplace, more violators are receiving tickets and fines in the mail. As a vehicle approaches a red light and surpasses a specific point under the street, it triggers the camera to take a photo of the vehicle crossing the line and in the middle of the intersection. Some cameras get images of the vehicle, the license plate, and even the driver' face. The system saves the following information to a database: Location of intersection Date and time of the incident Blown red lights account for roughly Speed of the vehicle 800 deaths Time lapse between red light and motorist crossing the intersection per year and up to $7 billion in property d amage. REAL LIFE: Motorists running a red light cause 22% The suspect of a drive-by purse snatching was caught using red light camera footage after dragging a 21-year-old student to her death because her hands were caught in the purse strings. OF ALL ACCIDENTS Human vision can't process and store images as fast as most vision systems. In the case of vehicle-related crimes, it can be almost impossible for a human to mentally capture important data like car make and model, license plate, driver description, etc. WINNER: VISION SYSTEM 3 MEDICAL IMAGING Breast Tomosynthesis adds a new dimension to mammograms and breast cancer detection. Also called 3D mammography, it allows radiologists to look inside the breast layer by layer, providing a clearer, more accurate view. Conventional 2D mammography gives a single, flattened image of the breast. During a 3D mammogram images are taken at multiple angles, which are combined to create a 3D rendering of the breast. ACCORDING TO AN OBSERVATIONAL STUDY OF BREAST REAL LIFE: TOMOSYNTHESIS IN ROUTINE SCREENING, THERE WAS: One woman received a "normal" or "negative" result from her conventional mammogram. She returned to the doctor several weeks later because she could 35% 53 INCREASE IN INCREASE IN feel a lump in her breast. An ultrasound detected the lump, but a review of her mammogram did not show the lump. On review of the tomosynthesis image, taken CANCER DETECTION INVASIVE CANCER RATES DETECTION RESULTS at the same time as part of a clinical study, the cancer could be clearly seen. 37% Medical imaging uses x-ray technology. which the human eye obviously doesn't have. Unless there is a visible lump, DECREASE IN DROP IN RECALL OR BIOPSY RATES there is no way for the human eye to CALLBACK RATES detect irregularities within the body. WINNER: VISION SYSTEM 4 PROD UCT SORTING AUTOMATION The recycling industry has gone even greener thanks to infrared (IR) spectroscopy. To provide more efficient sorting of recycled goods, IR spectroscopy is used to automate the sorting process. Different materials create different reflections, which are categorized into a range of infrared wavelengths. As materials move through the system, the reflections are read, analyzed, and sorted into the proper destination. The differences in light allow the computer to differentiate a variety of materials, including: CARDBOARD PLASTIC-COATED CARDBOARD DIFFERENT COLOR PLASTICS REAL LIFE: ECO plastics teamed up with TOMRA Sorting when they needed to provide a big independent bottler with a consistent supply of super-clean recycled PET (recycled polyethylene terephthalate). The purity levels needed by EC0 Plastics was significantly higher than industry standard. TOMRA Sorting went to work and developed a system using, among other things, IR spectrometry, to sort PET flakes to a higher standard of quality and purity than the industry had previously seen. It is almost impossible to tell the difference between a clear PVC and PET flake, even to a trained eye, so the human eye just doesn't stand a chance. WINNER: VISION SYSTEM 5 QUALITY AS SURANCE & PROD UCT SAFETY Infrared Thermography combines the approval of the human eye with the inspection of camera sensors, helping to minimize error and lower costs. In the f0od industry, controlling the temperature of perishable goods throughout production is essential. Automated thermography uses an IR camera and associated software to perform inspections without having to make contact with the item. It's used in a variety of applications, including food safety, equipment monitoring, quality assurance, and product inspection. IT CAN MONITOR TEMPERATURE, COLOR, AND SIZE. GENERAL USES IN THE FOOD INDUSTRY INCLUDE: MEASURING FOOD TEMPERATURE (MARGIN OF ERROR +/-2°C) MONITORING CO OKING EQUIPMENT PACKAGE INSPECTION REAL LIFE: In the summer of 2012, manufacturer Buona Vita had to recall almost 400,000 Ibs of frozen, ready-to-eat meat and poultry products for possible Listeria contamination. This accounted for millions of dollars in lost revenue by the time the situation was resolved. Listeria grows in cooler temperatures, so ensuring food is thoroughly pre-cooked to proper temperatures, as well as handled by strict sanitation guidelines, should remove any chance of contamination. While the quality inspection process of Buona Vita was not openly discussed during the incident, strict product inspection systems that include IR thermography may have removed the need for a recalI. While the human eye can certainly do well at product inspection related to look, color, and size of the product, infrared vision for heat detection is a bit beyond us. Vision systems win again! WINNER: VISION SYSTEM SOURCES I. 10. http// 2. I. htp:// 3. 12. http://www.massgeneral.erg/imaging/services/30_mammography_tomosynthesis.aspr 4. 13. 5. 14. 6. 15. applications/intrared_imaging for_industrial applications 7. 16. 17. hmp:// 8. 9. http://auto.howstuftw 18. COPYRIGHT 2014 | VISIONONLINE.ORG | ALL RIGHT RESERVED

Human Eye vs Vision Systems

shared by Lupeer on Aug 04
The advantages and function of computerized vision systems.





Did you work on this visual? Claim credit!

Get a Quote

Embed Code

For hosted site:

Click the code to copy


Click the code to copy
Customize size