Computer Vision Using Machine
Learning
Lecture Slides
Acknowledgement
We thankfully acknowledge following Professors for using their
Lecture Slides, Tutorials, video lectures for preparing the course
materials
• Prof. Andrew Ng, Stanford University
• Prof. Fei Fei Li, Stanford Vision Lab
• Prof. Mubarak Shah, UCF
• Prof. Geoffrey E. Hinton, University of Toronto
• Prof. P K Biswas, IIT Kharagpur
and many more…
INTRODUCTION
Syllabus
• Introduction to Computer Vision: Image representation and filtering,
Interest point detection, SIFT, HOG, Object recognition /
categorization, Segmentation, Bag of words;
• Introduction to Machine Learning: Linear regression, Logistic
regression, Over-fitting and Regularization, Support Vector
Machine; Neural Network
• Convolution Neural Network: Architectures, Auto-encoders, Deep
Belief Network, Restricted Boltzmann Machine; Beyond image
classification: localization, detection, segmentation
• Recurrent Neural Networks and LSTM
• Computer vision problems using deep learning aka CNN , RNN etc.
Books
• Text Books:
– Image and video text recognition using convolutional neural networks:
Study of new CNNs architectures for binarization, segmentation and
recognition of text images, by Zohra Saidane, Publisher: LAP
LAMBERT Academic Publishing (April 5, 2011)
– Face Image Analysis with Convolutional Neural Networks, by Stefan
Duffner , Publisher: GRIN Verlag (October 18, 2013)
– Fundamentals of Deep Learning, Designing Next-Generation Artificial
Intelligence Algorithms By Nikhil Buduma, Publisher: O'Reilly Media
Books
• Reference Books/Papers:
– Neural Networks and Deep Learning, by Michael Nielsen, Online Book:
http://neuralnetworksanddeeplearning.com.
– ImageNet Classification with Deep Convolutional Neural Networks, Alex
Krizhevsky, Ilya Sutskever, Geoffrey E Hinton, NIPS 2012.
– Going Deeper with Convolutions, Christian Szegedy, Wei Liu, Yangqing Jia,
Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent
Vanhoucke, Andrew Rabinovich, 19-Sept-2014.
– Learning Hierarchical Features for Scene Labeling, Clement Farabet,
Camille Couprie, Laurent Najman and Yann LeCun, IEEE Transactions on
Pattern Analysis and Machine Intelligence, 2013.
– Learning Convolutional Feature Hierachies for Visual Recognition, Koray
Kavukcuoglu, Pierre Sermanet, Y-Lan Boureau, Karol Gregor, Michaël
Mathieu and Yann LeCun, Advances in Neural Information Processing Systems
(NIPS 2010), 23, 2010.
– Ciresan, Dan, Ueli Meier, and Jürgen Schmidhuber. “Multi-column deep
neural networks for image classification.” Computer Vision and Pattern
Recognition (CVPR), 2012 IEEE Conference on. IEEE, 2012.
Grades
• Mid-Semester
– Percentage of grade : 20%
• End-Semester
– Percentage of grade : 30%
• Assignments:
– Percentage of grade : 50%
• Attendance below 75% --- F Grade
• Random Attendance
What is (computer) vision?
• When we “see” something, what does it mean?
• An color image, it is just an matrix of RGB values
(color pixels)
• Semantic definition of the image by computer
Task
Interpretation:
A Bird / blue sky and green land /
clouds / sign of peace
Why important now?
• In every 60 seconds, approximately 150 hours video
data are uploaded
• For a human observer, its almost impossible to
annotate such large amount of data
• Intelligent machine vision such as automatic image
or video understanding is essential
• A large space of applications
Application Areas
Field of study that gives computers the
ability to learn without being explicitly
programmed.
By: Arthur Samuel (1959)
Applications
• Database mining:
– Web click data, medical records, biology, engineering etc.
• Application that can’t be programmed by hand
– Autonomous helicopter, handwriting recognition, NLP,
computer vision
• Self Customizing Programs Recommendation
– Amazon, Netflix
• Understanding human learning
– Brain, real AI
Learning Problem
• Well-posed learning problem:
A computer program is said to learn from
experience E with respect to some task T and some
performance measure P, if its performance on T, as
measured by P, improves with experience E.
– By Tome Mitchell (1998)
Example
Your Email program watches which emails you do
or do not mark as a spam and based on that learns
how to better filter spam.
• Task (T): Classifying the emails as spam or not
• Experience (E): Watching you label emails as spam or not
spam
• Performance (P): The number of emails correctly classified
as spam / not spam
Machine Learning Algorithms
• Supervised Learning
• Un-supervised Learning
• Others: Reinforcement learning, recommender
system
• Where to apply which algorithms
Supervised Learning
Supervised Learning: Regression:
Right answer given Predict continuous valued output
Supervised Learning
Fruit : Apple or not?
Classification
Discrete values
1(Y) output (0 or 1)
Binary classifier
Multi-class classifier
0(N)
Fruit size
Supervised Learning
Fruit : Apple or not?
• Features:
Weight • color
• texture
• cost
• …..
Fruit size
Two features examples
Supervised Learning
x2
x1
Un-supervised Learning
No label
Clustering Algorithm
x2
x1
Application of Clustering Algorithms
• Organizing computer clusters
• Social network analysis
• Market segmentation
• Astronomical image/data analysis
• Speaker recognition and many more…
