人工智能與模式識別的發(fā)展ppt課件_第1頁
人工智能與模式識別的發(fā)展ppt課件_第2頁
人工智能與模式識別的發(fā)展ppt課件_第3頁
人工智能與模式識別的發(fā)展ppt課件_第4頁
人工智能與模式識別的發(fā)展ppt課件_第5頁
已閱讀5頁,還剩72頁未讀, 繼續(xù)免費閱讀

下載本文檔

版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進行舉報或認領(lǐng)

文檔簡介

1、1 1Some Recent Development of Intelligent PR and ApplicationsGuanghui He 2 2What are Biometrics?Biometrics are automated methods of recognizing a person based on the acquired physiological or behavioral characteristics Finger Scan52.1%Keystroke Scan0.3%Facial Scan11.4%Hand Scan10.0% Othe

2、rs12.4%Iris Scan7.3%Voice Scan4.1%Signature Scan2.4%Percentage of usage (Source: International biometric group)33A ScenarioTwo Al Qaeda(“基地組織)基地組織) suspects were recently taken into custody by U.S. immigration authorities as they tried to enter the United States after their fingerprints were matched

3、 with ones lifted by U.S. military officials from documents found in caves in Afghanistan阿富汗)阿富汗). Why Biometric Technologies?For Security Reasons4 4Example 1: SFinGe - Synthetic Fingerprint Generator developed at the Biometric Systems Lab,University of Bologna ITALY, is utilized to: compare differe

4、nt fingerprint matching algorithms train pattern recognition techniques that require large learning-sets (e.g. neural network) easily generate a large number of “virtual users” to develop and test medium/large-scale fingerprint-based systems 5 53-D model (pressure in on-line model) Modeling by defor

5、mation Modeling segments (conics, splines) Example 2: generation of synthetic signature Assembling (desegmentation) of 2-D model6 6 Example 3: Privacy protection: After enrollment, a true object (e.g. image of face, fingerprint or voice signal) is intentionally distorted using irreversible transform

6、 - Cancelable biometrics (Ratha, Connell, Bolle, 2019) Skin distortion (fingerprint) (source: Biometric Systems Lab, University of Bologna)Face image is warped with bilinear interpolation (source: Serif Inc.) Some More Examples: Generation of synthesis fingerprints Generation of synthetic signatures

7、 (handwriting modeling is a relevant problem) Iris recognition and synthesis Information fusion in biometrics Speech-to-animated-face78Where do we need biometrics? Traditional application: human identification Recent advances: Early warning paradigm Designing simulators for HQP training systems Sens

8、ing in robotics9Early detection and warningSemantic domainIndividualBiometric sensorSignal processingDecision makingRaw biometric dataBasic configurationFeature spaceApplication: physical access control systemSensorsExtractorsImage- andsignal processingalgorithmClassifiersBiometricsVoice, signature,

9、 face, fingerprint, iris, hand geometry, etcData Rep.Audio signal, image, infrared imageFeatureVectorsScoresDecision:Match, Non-match,InconclusiveBiometric databasesLevel 1: document-checkDatabases (Watch-list)Level 2: biometrics1011Laboratory experiments12Early warning system components:- Supports

10、facial analysis Skin temperature evaluation Detection of disguise: wig and other artificial materials, and surgical alternations Evaluation of blood vessel flow (modeling expressions) Other physiological / medical measurements (alcohol / drug abuse)Infrared biometrics and decision supportMid-infrare

11、d: 3-5 m, far-infrared: 8-12mTemperature value 32.8754 0C is detected in a point 13Early warning system components:Blood flow rate analysis (from infrared)Visualization of the blood flow rate from the upper rectangle of (a)Thermal image of subject at the beginning of answering the question “Do you h

12、ave that stolen $20 on you right now?”Thermal image of subject at the end of answering the questionVisualization of the blood flow rate from (b). The difference is significant (from I. Pavlidis report)14Early warning system: decision-makingAlternative mAlternative 2insufficiency of informationINDIVI

13、DUALbiometricsDegrees of beliefBiometric sensorTEMPORAL faults of biometric sensorserrors of biometric sensorsMass assignmentsAlternative 1Belief functionUpdatingDecision making in semantic form15Early warning security access control system:Semantic processorGait-biometric processorGait features pro

14、cessorThe ground reaction forceGenderPregnancyFatigueInjuriesAfflictionsDrunkennessGround reaction force processorDiscriminative gait biometric in semantic formGait biometrics analysis and decision-making assistance16Face capturingFitting points0001001001010011010010010010010110010010001000010010110

15、100100101001001001000 File (mesh/colour)3D Face modelEarly warning system components:17Face capturingFitting points0001001001010011010010010010010110010010001000010010110100100101001001001000 File (mesh/colour)3D Face modelEarly warning system components:18Other applications:Biometric data modeling

16、for HQP trainingProcessing of screened dataProcessing of pre-screeneddataDialogsupportDecision-making supportVisible band camera IR band camera Synthetic image of an individualVoice analyzer Officer-in-training 19Perspectives: humanoid robotsSensing in roboticsRobot head developed by Dr. Marek Perko

17、wskiat Portland State University Emotion synthesis Robot speech2020Its Similarity and Pattern Matching!What is Measurement ?Just a Comics Joke? No! More Than That2121Pattern RecognitionnCognition (Learning)nRe-CognitionnClassificationnIdentificationnVerificationnClustering22223D Object Recognition23

18、23Table of ContentsnBACKGROUNDnTHEORYnEXPERIMENTS and ILLUSTRATIONSnFUTURE RESEARCH2424Linear CombinationnObject 1 A1nObject 2 A2nObject 3 A3nObject 4 A4nObject A4= a A1+ bA2 +cA3 +d25253D Recognition BackgroundWidely usedindustrial parts inspection military target identificationCAM/CAD engineering

19、design image/vision understanding, interpretation, visualization, and recognition26263D Recognition BackgroundRecognition 3D objectsRigid Objects Fixed shapesDeformable Objects Variable shapesArticulated Objects Fewer methods proposed Brooks ACRONYM system using symbolic reasoning. Grimson et al ext

20、ended the interpretation of tree approach to deal with 2-D objects with articulated components 27273D Recognition BackgroundnExtended Linear Combination Method (LC)nSimpler preprocessing nSimpler and faster computation nApplicable to many articulated object recognition, understanding, interpretation

21、, and visualization2828THEORYnExtended Linear Combination Method (LC)nbased on the observation that novel views of objects can be expressed as linear combination of the stored views (from learning)n It identifies objects by constructing custom-tailored templates from stored two-dimensional image mod

22、els. 2929Linear CombinationModelan image consists of a list of feature points observed in the image 3030Linear CombinationRecognition: An unknown object is matched with a model by comparing the points in an image of the unknown object with a template-like collection of points produced from the model

23、31313232333334343535Experitment-1Match same objects3636Experiment-1 Result3737Experiment-23838Experiment-3 3939Experiment-3 Result4040Experiment-44141Experiment-4 ResultRejectedRejected Too42424343444445454646Color Biometric Imaging Analysis4747Items to be discussed:nClustering and K-means algorithm

24、nStatisticalnUnsupervisednColor Representation and Color Image Segmentation4848Supervised Classification and minimum distance classification nMinimum Distance ClassificationnSupervisednFind the center of known patterns of each class nClassify unknown patterns into the class that is “closest” to it.n

25、Ci xiixN14949Color Image Segmentation: Hue Component C1 Green Yellow 1 Red (H1) C2 Blue Magenta (H2) 5050Color Image SegmentationnTask:nStudy the K-means algorithm in hue space.nInteresting: nPeriodical Circular Property of hue componentnnew Measure of Distance.nProblem:nK-means algorithm is based o

26、n the measure of distance and definition of center5151Hue Component ClusteringnDefinition 1: Distance of Hue ValuesnDefinition 2: Directed Distance of Hue ValuesnTricky: Addition of Directed DistancenDefinition 3: Interval and Its Midpoint in H Space.nDefinition 4: Center of a Set of Points in Hue S

27、pacenTheory: Euclidean Theory of Center in Hue Space5252Hue Component ClusteringnDefinition 1: Distance of Hue Values2121212121 2),(HHHHHHHHHHd5353Hue Component ClusteringnDefinition 2: Directed Distance of Hue ValuesnTricky: Addition of Directed Distancenthe following vector addition property no lo

28、nger holds: 211212121221121221,)(22),(HHHHHHHHHHHHHHHHHHd),(),(),(322131HHdHHdHHd5454Hue Component ClusteringnRevisit definition: Interval and Its Midpoint in H Space.nRevisit definition : Center of a Set of Points in Hue SpacenRevisit the Proof of Theory: Euclidean Theory of Center in Hue Space5555

29、Color Image SegmentationnI and H components are of Interest. nGood color image segmentation algorithms should consider and combine bothnVariation of light intensity and occlusion:n hue component is betternColor information is lost:nIntensity component is better nFuzzy member function is introduced5656Color Image Segmentation - Experiment 1Intensity Distinguishable(a) Origi

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
  • 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負責(zé)。
  • 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準確性、安全性和完整性, 同時也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

最新文檔

評論

0/150

提交評論