Introduction to Pattern RecognitionWhat is a Pattern?Examples of PatternsPR DefinitionsExample Problem:Handwritten Digit RecognitionRole of Machine LearningMachine LearningClassification Process(Decision as opposed to Inference)Pattern Recognition ApplicationsPattern Recognition ProcessesGeneralizationPattern Recognition SystemDesign CycleDocument Recognition ApplicationsAddress Interpretation ProblemContextual InformationCSE 555: Sargur Srihari 1Introduction to Pattern RecognitionSargur N. [email protected]. of Computer Science & EngineeringState University of New York at BuffaloCSE 555: Sargur Srihari 2What is a Pattern?“A pattern is the opposite of chaos; it is an entity vaguely defined, that could be given a name.”A pattern is an abstract object, such as a set of measurements describing a physical object.CSE 555: Sargur Srihari 3Examples of PatternsHandwritten CharactersPostnet Bar CodeFingerprintUPC BarCodeAnimal FootprintData TrendCSE 555: Sargur Srihari 4PR Definitions• Theory, Algorithms, Systems to put Patterns into Categories• Classification of Noisy or Complex Data• Relate Perceived Pattern to Previously Perceived PatternsCSE 555: Sargur Srihari 5Example Problem:Handwritten Digit Recognition• Handcrafted rules will result in large no of rules and exceptions• Better to have a machine that learns from a large training setWide variability of same numeralCSE 555: Sargur Srihari 6Role of Machine Learning• Principled way of building high performance information processing systems•ML vsPR– ML has origins in Computer Science– PR has origins in Engineering– They are different facets of the same field• Language Related Technologies– IR, NLP, DAR, ASR– Humans perform them well– Difficult to specify algorithmicallyCSE 555: Sargur Srihari 7Machine Learning• Programming computers to use example data or past experience• Well-Posed Learning Problems– A computer program is said to learn from experience E – with respect to class of tasks T and performance measure P, – if its performance at tasks T, as measured by P, improves with experience E.CSE 555: Sargur Srihari 8Classification Process(Decision as opposed to Inference)CSE 555: Sargur Srihari 9Pattern Recognition Applications•SENSORY• Vision– Face/Handwriting/Hand• Speech – Speaker/Speech• Touch– Haptics• Olfaction– Apple Ripe?• TEXTUAL DATA• Text Categorization• Information Retrieval• Data Mining• Intrusion Detection• Genome Sequence MatchingCSE 555: Sargur Srihari 10Pattern Recognition Processes• Objects to be classified are sensed by transducer (camera)• Signals are preprocessed• Features are extracted• Classification is emittedCSE 555: Sargur Srihari 11GeneralizationCSE 555: Sargur Srihari 12Pattern Recognition SystemCSE 555: Sargur Srihari 13Design CycleCSE 555: Sargur Srihari 14Document Recognition Applications• Optical Character Recognition(OCR)• Handwriting Recognition• Writer RecognitionCSE 555: Sargur Srihari 15Writer RecognitionPreprocessingFeaturesSimilarityCSE 555: Sargur Srihari 16Address Interpretation ProblemPattern recognition tasks– object recognition (address vs non-address)– two-class discrimination (mp vs hw)– few class recognition (digits)– holistic vs analytical (words)– contextual-hmm(zip codes, words)– Many classes, but cataloged (postal directory)Contextual Information• Country/State/City• ZIP Code• Street Name• Primary No (Street/PO Box )• Secondary No (Apt)• Firm/Personal NameCSE 555: Sargur Srihari