●●● ●●●● ●●●●● ●●●● ●●●●● Face Detection ●●●● ●●●● o Neural network-Based algorithm The basic algorithm used for face detection nput image pyramid Extracted window Corrected lighting Histogram equalized Receptive fields 20 by 20 pixels) Hidden units Network Output Qo 20 by Neural network Extraction Techniques
Face Detection ⚫ Neural Network-Based Algorithm ⚫ The basic algorithm used for face detection Extraction Techniques
●●● ●●●● ●●●●● ●●●● ●●●●● Face Detection ●●●● ● Face Recognition Facial Expression Analysis ● Enrich the Xml o easier for user to search the content of video Extraction Techniques
Face Detection ⚫ Face Recognition ⚫ Facial Expression Analysis ⚫ Enrich the XML ⚫ Easier for user to search the content of video Extraction Techniques
●●● ●●●● ●●●●● ●●●● ●●●●● Speech Recognition ●●●● Speech recognition technology can make any spoken data useful for library indexing and retrieval SPEECH COGNITION SPEECH-AVVARE USER M CROPHONE SOUND CARD ENGINE APPLICATION WAATD ythat tei量? TAYM IH S WHAT TIME IH IT IS IT User speaks into Microphone captures Sound card recognition Application Sound waves and ne coverts processes generates electrical acoustical sign signal to words as: text to digital signal input words Extraction Techniques
Speech Recognition Speech recognition technology can make any spoken data useful for library indexing and retrieval Extraction Techniques
●●● ●●●● ●●●●● ●●●● ●●●●● Speech Recognition Engine ●●●● Acoustic models Input signal Language model 515253 P(xx ↓x o pyx) Feature extraction p(zx) Ao p(xly lyly) p(zy) Lexicon p(×2) ba=b同a p(y2) p(z2) 卹p=团l arch-engine W= argmax f(ylw,)P(w) y151) Recognised word sequence Extraction Techniques
Speech Recognition Engine Extraction Techniques