2. • Computer Vision
• Natural Language Processing Applications
• Autonomous Vehicle
• Healthcare
• Retail
• Manufacturing
• Agriculture
• Smart City
• Government
4. • Computer vision (CV) is the object recognition.
• It teaches the machine how to recognize the objects and react differently
depending on the object classes.
• Compared with the traditional computer vision algorithmic approach, it is difficult
to cover all features extraction using an algorithmic approach because similar
objects may have different features.
• Deep convolution neural network (DCNN) does not require defining the features
for object recognition. DCNN automatically updates the model parameters for
feature extraction during the training.
5. o Object recognition is further divided into image classification, object localization, and
object detection.
o The image classification recognizes the objects in the image, and then assigns the class
labels to the objects.
o The object localization locates objects in the image and draws a bounding box around
the objects.
o Object detection combines both image classification and object localization. It can
detect multiple objects in the image, estimate their size and position, and then draw the
bounding box around the objects.
6. • The object detection can also generate the pixel-wise mask over the object to
determine the exact object shape and called instance segmentation. It can further
label each pixel in the image including the background and is called semantic
segmentation.
7. o Gesture recognition in the retail market is expected to
grow by 27.54% from 2018 to 2023.
o The gesture-based interface allows the user to control
different devices using hand or body motion.
o It first captures the hand or body motion through the
camera and then analyzes the motion in each frame.
Through the machine learning model, it recognizes
different motions and identifies the gesture, followed by
the corresponding actions.
o It is widely applied to home automation, and people can
control home devices through gestures, activate the light,
adjust the room temperature, and turn on/off the music
system.
o It can recognize individual family members and perform
predefined actions based on personal history.
8. • Medical Diagnosis
• Medical imaging (ultrasound, X-rays, computerized tomography scan – CT scan, and
magnetic resonance imaging – MRI) is a useful cancer diagnosis tool.
• Retail Applications
• Retail also benefits from machine vision, which teaches the machine to recognize the
items in the images and videos. It helps the retail to sell the products with better inventory
control. Currently, 43% of retail artificial intelligence applications are related to machine
vision.
• Airport Security
• Facial recognition is an important airport security application, especially for passenger
processing. It quickly recognizes and validates the passenger’s identity and alerts the
airport authority for security violations.
16. o Telemedicine: If the patients live in rural areas, they are geographically isolated from the healthcare facilities
(clinic, laboratory, hospital). Artificial intelligence can solve problems and improve the quality of life.
o Medical Imaging
17. • Smart Medical
Device
• Electronic Health
Record
• Drug Development
• Clinical Trial
• Medical Robotics
• Elderly Care
26. • Human Service
• Law Enforcement
• Homeland Security
• Legislation
• Ethics
• Public Perspective
27. • Human Service
• Law Enforcement
• Homeland Security
• Legislation
• Ethics
• Public Perspective
28. 1. What are the Artificial intelligence applications?
29. • Albert Chun-Chen Liu, Oscar Ming Kin Law, Iain Law - Understanding Artificial
Intelligence_ Fundamentals and Applications-Wiley-IEEE Press (2022).