Pattern Recognition


Published on

  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Pattern Recognition

  1. 1. PATTERN RECOGNITION Talal A. Alsubaie SFDA
  2. 2. OUTLINES <ul><li>What is a pattern? </li></ul><ul><li>What is A pattern Class ? </li></ul><ul><li>What is pattern recognition? </li></ul><ul><li>Human Perception </li></ul><ul><li>Examples of applications </li></ul><ul><li>The Statistical Way </li></ul><ul><li>Human and Machine Perception </li></ul><ul><li>P attern R ecognition </li></ul><ul><li>Pattern Recognition Process </li></ul><ul><li>Case Study </li></ul>
  3. 3. WHAT IS A PATTERN? <ul><li>A pattern is an abstract object, or a set of measurements describing a physical object. </li></ul>
  4. 4. WHAT IS A PATTERN CLASS ? <ul><li>A pattern class (or category) is a set of patterns sharing common attributes. </li></ul><ul><li>A collection of “similar” (not necessarily identical) objects. </li></ul><ul><li>During recognition given objects are assigned to prescribed classes. </li></ul>
  5. 5. WHAT IS PATTERN RECOGNITION? <ul><li>Theory, Algorithms, Systems to put Patterns into Categories </li></ul><ul><li>Relate Perceived Pattern to Previously Perceived Patterns </li></ul><ul><li>Learn to distinguish patterns of interest from their background </li></ul>
  6. 6. HUMAN PERCEPTION <ul><li>Humans have developed highly sophisticated skills for sensing their environment and taking actions according to what they observe, e.g., </li></ul><ul><ul><li>Recognizing a face. </li></ul></ul><ul><ul><li>Understanding spoken words. </li></ul></ul><ul><ul><li>Reading handwriting. </li></ul></ul><ul><ul><li>Distinguishing fresh food from its smell. </li></ul></ul><ul><li>We would like to give similar capabilities to machines. </li></ul>
  9. 9. GRID BY GRID COMPARISON Grid by Grid Comparison A A B
  10. 10. GRID BY GRID COMPARISON 0 0 1 0 0 0 1 0 0 1 1 1 1 0 0 1 1 0 0 1 No of Mismatch= 3 A A B 0 1 1 0 0 1 1 0 0 1 1 0 1 0 0 1 1 0 0 1
  11. 11. GRID BY GRID COMPARISON Grid by Grid Comparison A A B
  12. 12. GRID BY GRID COMPARISON 0 0 1 0 0 0 1 0 0 1 1 1 1 0 0 1 1 0 0 1 1 1 1 0 0 1 0 1 0 1 1 1 0 1 0 1 1 1 1 0 No of Mismatch= 9 A A B
  13. 13. PROBLEM WITH GRID BY GRID COMPARISON <ul><li>Time to recognize a pattern - Proportional to the number of stored patterns ( Too costly with the increase of number of patterns stored ) </li></ul>Solution Artificial Intelligence A-Z a-z 0-9 */-+1@#
  14. 14. HUMAN AND MACHINE PERCEPTION <ul><li>We are often influenced by the knowledge of how patterns are modeled and recognized in nature when we develop pattern recognition algorithms. </li></ul><ul><li>Research on machine perception also helps us gain deeper understanding and appreciation for pattern recognition systems in nature. </li></ul><ul><li>Yet, we also apply many techniques that are purely numerical and do not have any correspondence in natural systems. </li></ul>
  15. 15. PATTERN RECOGNITION <ul><li>Two Phase : Learning and Detection . </li></ul><ul><li>Time to learn is higher. </li></ul><ul><ul><li>Driving a car </li></ul></ul><ul><li>Difficult to learn but once learnt it becomes natural . </li></ul><ul><li>Can use AI learning methodologies such as: </li></ul><ul><ul><li>Neural Network. </li></ul></ul><ul><ul><li>Machine Learning. </li></ul></ul>
  16. 16. LEARNING <ul><li>How can machine learn the rule from data? </li></ul><ul><ul><li>Supervised learning : a teacher provides a category label or cost for each pattern in the training set. </li></ul></ul><ul><ul><li>Unsupervised learning : the system forms clusters or natural groupings of the input patterns. </li></ul></ul>
  17. 17. <ul><li>Classification (known categories) </li></ul><ul><li>Clustering (creation of new categories) </li></ul>CLASSIFICATION VS. CLUSTERING Category “A” Category “B” Clustering (Unsupervised Classification) Classification (Supervised Classification)
  18. 18. PATTERN RECOGNITION PROCESS (CONT.) Post- processing Classification Feature Extraction Segmentation Sensing input Decision
  19. 19. PATTERN RECOGNITION PROCESS <ul><li>Data acquisition and sensing: </li></ul><ul><ul><li>Measurements of physical variables. </li></ul></ul><ul><ul><li>Important issues: bandwidth, resolution , etc. </li></ul></ul><ul><li>Pre-processing: </li></ul><ul><ul><li>Removal of noise in data. </li></ul></ul><ul><ul><li>Isolation of patterns of interest from the background. </li></ul></ul><ul><li>Feature extraction: </li></ul><ul><ul><li>Finding a new representation in terms of features. </li></ul></ul><ul><li>Classification </li></ul><ul><ul><li>Using features and learned models to assign a pattern to a category. </li></ul></ul><ul><li>Post-processing </li></ul><ul><ul><li>Evaluation of confidence in decisions. </li></ul></ul>
  20. 20. CASE STUDY <ul><li>Fish Classification: </li></ul><ul><ul><li>Sea Bass / Salmon. </li></ul></ul><ul><li>Problem : Sorting incoming fish </li></ul><ul><li>on a conveyor belt according to </li></ul><ul><li>species. </li></ul><ul><li>Assume that we have only two kinds of fish: </li></ul><ul><ul><li>Sea bass. </li></ul></ul><ul><ul><li>Salmon. </li></ul></ul>Salmon Sea-bass
  21. 21. CASE STUDY (CONT.) <ul><li>What can cause problems during sensing? </li></ul><ul><ul><li>Lighting conditions. </li></ul></ul><ul><ul><li>Position of fish on the conveyor belt. </li></ul></ul><ul><ul><li>Camera noise. </li></ul></ul><ul><ul><li>etc… </li></ul></ul><ul><li>What are the steps in the process? </li></ul><ul><ul><li>Capture image. </li></ul></ul><ul><ul><li>Isolate fish </li></ul></ul><ul><ul><li>Take measurements </li></ul></ul><ul><ul><li>Make decision </li></ul></ul>
  22. 22. CASE STUDY (CONT.) Classification Feature Extraction Pre-processing “ Sea Bass” “ Salmon”
  23. 23. CASE STUDY (CONT.) <ul><li>Pre-Processing: </li></ul><ul><ul><li>Image enhancement </li></ul></ul><ul><ul><li>Separating touching or occluding fish. </li></ul></ul><ul><ul><li>Finding the boundary of the fish. </li></ul></ul>
  24. 24. HOW TO SEPARATE SEA BASS FROM SALMON? <ul><li>Possible features to be used: </li></ul><ul><ul><li>Length </li></ul></ul><ul><ul><li>Lightness </li></ul></ul><ul><ul><li>Width </li></ul></ul><ul><ul><li>Number and shape of fins </li></ul></ul><ul><ul><li>Position of the mouth </li></ul></ul><ul><ul><li>Etc … </li></ul></ul><ul><li>Assume a fisherman told us that a “sea bass” is generally longer than a “salmon”. </li></ul><ul><li>Even though “sea bass” is longer than “salmon” on the average, there are many examples of fish where this observation does not hold. </li></ul>
  25. 25. HOW TO SEPARATE SEA BASS FROM SALMON? <ul><li>To improve recognition, we might have to use more than one feature at a time. </li></ul><ul><ul><li>Single features might not yield the best performance. </li></ul></ul><ul><ul><li>Combinations of features might yield better performance. </li></ul></ul>
  26. 26. FEATURE SELECTION “ Good” features “ Bad” features
  28. 28. DECISION BOUNDARY (CONT.) More complex model result more complex boundary
  29. 29. DECISION BOUNDARY (CONT.) Different criteria lead to different decision boundaries
  30. 30. DECISION BOUNDARY (CONT.) <ul><li>What if a customers find “ Sea bass ” in there “ Salmon ” can? </li></ul><ul><li>We should also consider costs of different errors we make in our decisions. </li></ul>
  31. 31. DECISION BOUNDARY (CONT.) <ul><li>For example, if the fish packing company knows that: </li></ul><ul><ul><li>Customers who buy salmon will object vigorously if they see sea bass in their cans. </li></ul></ul><ul><ul><li>Customers who buy sea bass will not be unhappy if they occasionally see some expensive salmon in their cans. </li></ul></ul><ul><li>How does this knowledge affect our decision? </li></ul>
  32. 32. CASE STUDY (CONT.) <ul><li>Issues with feature extraction: </li></ul><ul><ul><li>Correlated features do not necessary improve performance. </li></ul></ul><ul><ul><li>It might be difficult to extract certain features. </li></ul></ul><ul><ul><li>It might be computationally expensive to extract many features. </li></ul></ul><ul><ul><li>Missing Features. </li></ul></ul><ul><ul><li>Domain Knowledge. </li></ul></ul>
  34. 34. DEMO
  35. 35. DEMO <ul><li>Online face detector demo: </li></ul><ul><ul><li> </li></ul></ul>
  36. 36.
  37. 37. DEMO (CONT.) <ul><li>With my friend “ Albert Einstein ” </li></ul>
  38. 38. VIDEO DEMO
  39. 39. Q & A
  40. 40. THANK YOU