l21_int_prob_gra

l21_int_prob_gra - Visual Interpretation using...

Info iconThis preview shows pages 1–13. Sign up to view the full content.

View Full Document Right Arrow Icon
Visual Interpretation using Probabilistic Grammars Paul Robertson
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Model-Based Vision • What do the models look like • Where do the models come from • How are the models utilized 2
Background image of page 2
The Problem 3
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4
Background image of page 4
5
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Optimization/Search Problem Find the most likely interpretation of the image contents that: 1. Identifies the component parts of the image correctly. 2. Identifies the scene type. 3. Identifies structural relationships between the parts of the image. Involves: Segmenting into parts, naming the parts, and relating the parts. 7
Background image of page 6
Outline Overview of statistical methods used in speech recognition and NLP Image Segmentation and Interpretation – image grammars – image grammar learning – algorithms for parsing patchwork images. 8
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Not any description – the best s s np vp vp np np pp np np noun noun verb noun verb noun prep noun swat flies like ants swat flies like ants Bad parse Good parse 9
Background image of page 8
What’s similar/different between image analysis and speech recognition/NLP? • Similar – An input signal must be processed. – Segmentation. – Identification of components. – Structural understanding. • Dissimilar – Text is a valid intermediate goal that separates Speech recognition and NLP. Line drawings are less obviously useful. – Structure in images has much more richness. 10
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Speech Recognition and NLP Speech Recognition Natural Language Processing Part of Segmentation Sentence speech into words tagging Parsing • Little backward flow • Stages done separately. • Similar techniques work well in each of these phases. • A parallel view can also be applied to image analysis. 11
Background image of page 10
Speech Understanding • Goal: Translate the input signal into a sequence of words. – Segment the signal into a sequence of samples. •A = a 1 , a 2 , . .., a m a i – Find the best words that correspond to the samples based on: • An acoustic model. – Signal Processing – Prototype storage and comparator (identification) • A language model. •W = w 1 , w 2 , . .., w m w i –W opt = arg max w P(W|A) –W opt =argmax w P(A|W) P(W)
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
n ( ( W P ) = w P i | w 1 ,..., w i 1 ) i = 1 n ( ( W P ) = w P i | Φ ( w 1 ,..., w i 1 )) i = 1 n ( ( W P ) = w P i | Φ i 1 ) i = 1 ( w P i | w i 1 , w i 2 ) = f ( w i | w i 1 , w i 2 ) ( w P i | w i 1 , w i 2 ) = λ 3 f ( w i | w i 1 , w i 2 ) + λ 2 f ( w i | w i 1 ) + λ 1 f ( w i ) λ 1 + λ 2 + λ 3 = 1 Using the above P(W) can be represented as a HMM and solved efficiently using the Viterbi algorithm. The good weights
Background image of page 12
Image of page 13
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 41

l21_int_prob_gra - Visual Interpretation using...

This preview shows document pages 1 - 13. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online