4. Introduction | Motivation
Makes computer vision a possibility, hence
enhancing power of Artificial Intelligence.
There is significant interest in creating light
weight and mobile systems that can identify
objects using vision
Numerous practical application makes
Image Recognition a motivating field of study.
5. Introduction | Motivation
“ On an average,
over 300 million
images are
uploaded on
Facebook daily “
Source: Zephoria
6. Introduction | Motivation
“ The image
recognition market is
estimated to grow
from US$16 Billion
in 2016 to US$39
Billion by 2021 ”
Source: Zephoria
7. Introduction | What is Image Recognition?
• Image recognition is the process of
identifying and detecting an object or
a feature in a Digital Image.
• It is also known as Computer Vision.
8. Introduction | What is a digital image?
• A digital image is a representation
of a 2D image using a finite set of
digital values for each pixel.
• A pixel is the smallest
independent block of a digital
image.
• The digital values of these pixels
are processed and used in Image
Recognition and in other areas of
Image Processing.
10. Introduction | Steps in Image Recognition
Data acquisition and sensing
Preprocessing
• Removal of noise
• Isolation of patterns of interest from
the background (Segmentation)
Feature Extraction
• Finding a new representation in terms
of features (Detection)
11. Introduction | Steps in Image Recognition
Model Learning and Estimation
• Learning a mapping between features and
pattern groups.
Classification
• Using learned models to assign a pattern to
a predefined category
Post processing
• Evaluation of confidence in decisions.
• Exploitation of context to improve
performances.
12. Introduction | Preprocessing
• Images are preprocessed
to be fed as input into the
algorithm.
• Preprocessing helps in
better feature extraction
from the image.
13. Introduction | Edge Detection
Common methods of Edge Detection:
• Canny Edge Detection: Uses
calculus of variations (most widely
used) – optimizes a given functional
• Sobel Edge Detection: It is a
discrete differentiation operator,
computing an approximation of the
gradient of the image intensity
function
14. Introduction | Practical Applications of Image
Recognition
Medical Imaging
• extensively used for cancer detection, retinopathy
detection, improving quality of imperfect images.
Industrial Applications
• fault detection in manufacturing
Commercial Applications
• In store shopper tracking
• Inventory control
15. Introduction | Practical Applications of Image
Recognition
Security
• Face & fingerprint & retina - iris recognition
Transportation
• Autonomous vehicles
Applications for creative media
• Deep dream
• Neural style transfer (prizma)
• Human and Computer interface
16. Introduction | Practical Applications of Image
Recognition
Geographic Information Systems
• Terrain Classification
• Meteorology
• Global inventory of human settlement
Astronomy
• Enhancement of telescopic images
• Recognition of astronomical bodies
• Eg: The Hubble Telescope
18. What is Deep Learning?
• Part of the machine learning field of
learning representations of data.
Exceptional effective at learning patterns.
• Utilizes learning algorithms that derive
meaning out of data by using a hierarchy of
multiple layers that mimic the neural
networks of our brain.
• If you provide the system tons of
information, it begins to understand it and
respond in useful ways.
19. Inspired by the Brain
• The first hierarchy of neurons that
receives information in the visual
cortex are sensitive to specific
edges while brain regions further
down the visual pipeline are sensitive
to more complex structures such as
faces.
• Our brain has lots of neurons
connected together and the strength
of the connections between neurons
represents long term knowledge.
22. Deep Learning | Architecture
A deep neural network
consists of a hierarchy of
layers, whereby each layer
transforms the input data
into more abstract
representations
e.g. edge -> nose -> face
The output layer combines
those features to make
predictions.
24. Deep Learning | More Layers -> Better Performance
The more
layers the
network has,
the higher-level
features it will
learn.
25. Deep Learning | Convolutional Neural Nets (CNN)
Convolutional Neural Networks learn a complex representation of visual data using vast
amounts of data.
• They are inspired by the human visual system and learn multiple layers of
transformations, which are applied on top of each other to extract a progressively more
sophisticated representation of the input.
E.g. Image is a 224*224*3 (RGB) cube and will be transformed to 1*1000 vector of probabilities.
41. KNIME | DeepLearning4J Extension
• Visually assemble networks using KNIME
nodes
• Integrates with other KNIME extensions
• e.g. Image Processing & Text Mining
• Networks can be trained and executed
on GPU and CPU
42. Deep learning benefits
Robust
• No need to design the features ahead of time - features are automatically
learned to be optimal for the task at hand
• Robustness to natural variations in the data is automatically learned
Generalizable
• The same neural net approach can be used for many different applications
and data types
Scalable
• Performance improves with more data, method is massively parallelizable