“Computer
vision is a field of artificial intelligence (AI) that uses machine learning
and neural networks to teach computers and systems to derive meaningful
information from digital images, videos and other visual inputs—and to make
recommendations or take actions when they see defects or issues.”
History of computer vision
Early
experiments in computer vision took place in the 1950s, using some of the first
neural networks to detect the edges of an object and to sort simple objects
into categories like circles and squares. In 1976 optical character recognition
(OCR) technology comes to recognize text printed in any font and intelligent
character recognition (ICR) that could read hand-written text that helped in document
reading, invoice processing, vehicle plate recognition, mobile payments,
machine conversion and other common applications. In 1982, neuroscientist David
Marr established algorithms for machines to detect edges, corners, curves, and
similar basic shapes. Later in the 1990s large sets of images were available
online for analysis, facial recognition programs that helped machines to
recognize specific people in photos and videos and used in today’s deep
learning. Because a system is trained to examine products and observe a
production asset, it can analyze thousands of products or processes in minutes,
noticing very small/invisible that surpass human abilities. In today’s world
Computer vision industries range from energy, utilities, manufacturing,
automotive, medical, public safety division, and the market is continuing to
grow.
How computer vision works:
Computer
vision is all about pattern recognition and it got trained by feeding lots of
images that have been labeled and subject to various software algorithms, that
allow the computer to track down patterns like color, shape etc. that relates
to those labels. Here are a few steps.
Step #1:
Image acquisition:
The digital image of a camera where any 2D or 3D camera or sensor can be used
to provide image frames.
Step #2: Pre-processing:
Pre-processing includes noise reduction, contrast enhancement, re-scaling, or
image cropping to get it ready for processing.
Step #3: Computer vision algorithm/processing/segmentation:
“The image processing algorithm, most popularly a deep learning model (DL
model), performs image recognition, object detection, image segmentation, and
classification on every image or video frame.”
Step #4: Automation logic to recognition: The AI algorithm output information needs to be processed with conditional rules set based on the use case. And performs automation on input data and create output. “For example, pass or fail for automatic inspection applications, match or no-match in recognition systems, flag for human review in insurance, surveillance and security, military, or medical recognition applications”.
Key applications of Computer vision
Self-Driving
Cars: Computer vision
enables self-driving cars to make sense of their surroundings by processing
images from cars camera in real-time to find road terrific signs, detect other
cars, objects, pedestrians etc. to drive safely.
Healthcare: It helps automate tasks like finding human
fall detection in X-rays, MRI scans and detecting cancerous moles in skin
images to risk and trigger alerts.
Security/Facial
recognition: Computer
vision helps match the people’s images to their identities by comparing facial
features with database of face profiles. In useful in video surveillance and
security to control unlawful activities.
Super-resolution imaging (SR): it helps interpret numerous low-quality images differently for their unique information and based on the images variations this model produces a stream of high-quality images.
Augmented
Reality (AR):
Computer vision for augmented and virtual reality that enables smartphones,
tablets, and smart glasses to create immersive experiences by integrating
virtual surroundings in real-time. For example, computer vision algorithms can
help AR applications detect planes such as tabletops, walls and floors and
placing virtual objects in the physical world.
Optical
character recognition (OCR):
Computer vision helps in OCR to extract the text from images, scanned
documents, PDFs. Which is very in every industry.
Computer Vision in Accounting/Financial Industry
1) Document Processing and Entry: The financial sector depends on
document-intensive processes. Computer vision OCR helps in extracting relevant information
from images, PDFs, and other file formats and helps in automating data entry
and categorization, which can minimize manual errors and enhance operational
accuracy.
2) Customer Service: Computer vision contributes to
enhanced customer support by analyzing customer facial expressions and gestures
during video calls to scale customer satisfaction levels. This data advises
real-time adjustments to interactions, so customers feel satisfied.
3) Real-time Transaction Monitoring and
detecting any fraud or errors: Computer
Vision facilitates continuous monitoring of transactions to detect any unusual or
unauthorized access. Computer Vision-based systems constantly analyze
transaction data and cross-referencing it with historical patterns such as
unexpected large withdrawals, changes in spending behavior, suspicious
transfers, to identify potential fraudulent transactions and stop them immediately
to prevent financial losses.
4) Facial Recognition for Identity
Verification:
Computer Vision powered facial recognition technology which is considered as a
secure and convenient way for identity verification and eliminating the need
for manual verification processes.
Source:
https://www.ibm.com/topics/computer-vision
https://www.turing.com/kb/all-you-need-to-know-about-computer-vision
https://viso.ai/computer-vision/what-is-computer-vision/
https://www.linkedin.com/pulse/top-5-use-cases-computer-vision-financial
https://sjgrand.cn/integrating-computer-vision-accounting/
No comments:
Post a Comment