Probabilistic generative models for machine vision
- 1. Machine Vision Background Region Annotation by Hierarchical Region Topic Discovery Probabilistic Generative Models for Machine Vision Davide Bacciu bacciu@di.unipi.it Dipartimento di Informatica Università di Pisa Corso di Intelligenza Artificiale Prof. Alessandro Sperduti Università degli Studi di Padova A.A. 2008/2009 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 2. Machine Vision Background Region Annotation by Hierarchical Region Topic Discovery Outline 1 Machine Vision Background Introduction Visual Content Representation Machine Vision Applications 2 Region Annotation by Hierarchical Region Topic Discovery Visual Content Representation Hierarchical Probabilistic Latent Semantic Analysis Experimental Evaluation Davide Bacciu Probabilistic Generative Models for Machine Vision
- 3. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Introduction Visual content representation Visual features - how to identify and describe the single bits of visual information Image representation - how to describe the visual information within a picture Machine vision applications Scene and object recognition Image annotation Focus on Probabilistic Generative Models and the Bag of Words image representation Davide Bacciu Probabilistic Generative Models for Machine Vision
- 4. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Visual Features Descriptors How do we describe the single bits of visual information? Global descriptors Color histograms Shape features Texture Spatial Envelope Local descriptors Gabor filters Differential filters Distribution-based descriptors Davide Bacciu Probabilistic Generative Models for Machine Vision
- 5. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Scale Invariant Feature Transform (SIFT) Calculate gradient magnitude and orientation at a given keypoint (x, y ) and scale σ Compute 3D histogram of magnitude orientations in a 4 × 4 neighborhood of the keypoint Angles are quantized to 8 directions Robust to geometric and illumination distortion Davide Bacciu Probabilistic Generative Models for Machine Vision
- 6. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Visual Features Detectors How do we identify the interesting/informative parts of an image? Feature detectors Segment/Blob detectors Corner/Interest point detectors Affine invariant keypoint detectors Dense sampling Davide Bacciu Probabilistic Generative Models for Machine Vision
- 7. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Image Representation Straightforward for global descriptors An image corresponds to its descriptor Images are confronted, classified and indexed based on their global color and/or texture histograms Requires an additional step for local descriptors Local descriptors provide only information on a portion of the scene Probabilistic modeling of segments Bag of words Davide Bacciu Probabilistic Generative Models for Machine Vision
- 8. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Bag of Words Document Representation Count the occurrences of each dictionary word in your document Represent the document as a vector of word counts Definition (Bag of Words Assumption) Word order is not relevant for determining document semantics Davide Bacciu Probabilistic Generative Models for Machine Vision
- 9. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Bag of Words Image Representation Each image is a document and each visual patch is a word Count the occurrences of each dictionary visual word (visterm) in your image Represent the image as a vector of visterm counts (histogram) ... k−means Image Codebook Collection Local patches Vector Quantization Image Input Image BOW Collection Davide Bacciu Probabilistic Generative Models for Machine Vision
- 10. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Image Understanding Applications Images are typically represented as vector of features Discover the semantics of an image from its vectorial representation Object recognition (e.g. Airplane) Scene recognition (e.g. Forest) Image annotation (e.g. Sky, Tree, Boat, Sea, ...) Davide Bacciu Probabilistic Generative Models for Machine Vision
- 11. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Scene and Object Recognition Structured approaches Place Context Object/scene categories are represented as collections of connected parts characterized by Object Context distinctive appearance and spatial Part Context position Part-based probabilistic models Weakly structured approaches PixelContext Based on bag-of-word assumption Latent aspect models Davide Bacciu Probabilistic Generative Models for Machine Vision
- 12. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Latent Aspect Discovery Rooted into text processing Probabilistic models describing the generative process of image content using hidden (i.e. latent) variables Model likelihood is used to measure the fit of the unknown parameters with respect to the observations Model parameters are fitted by Expectation Maximization (EM) or variational methods Davide Bacciu Probabilistic Generative Models for Machine Vision
- 13. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Probabilistic Latent Semantic Analysis (pLSA) (I) P(d) P(z|d) P(w|z) d z w Wd D Observations are collected in the integer matrix n(wj , di ) Every observation is associated to a latent topic k by means of the hidden variable zk Each hidden topic k denotes a particular visual theme (e.g. an object) Davide Bacciu Probabilistic Generative Models for Machine Vision
- 14. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Probabilistic Latent Semantic Analysis (pLSA) (II) P(d) P(z|d) P(w|z) d z w Wd D pLSA is trained by Expectation Maximization (EM) Conditional independence is used to decompose model likelihood D W L(N |θ) = P(wj , di )n(wj ,di ) i=1 j=1 K P(wj , di ) = P(di )P(wj |di ) = P(di ) P(zk |di )P(wj |zk ) k =1 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 15. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications An Object Recognition Example Caltech-4 Dataset: cars, faces, motorbikes and airplanes on a cluttered background Fit a pLSA model with 7 latent topics (4 objects + 3 background) Determine the most likely topic of each visual word in image di P(zk |di )P(wj |zk ) P(zk |wj , di ) = K k =1 P(zk |di )P(wj |zk ) Figure: Images by Sivic et al 2005 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 16. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Latent Dirichlet Allocation (LDA) β φ K α θ z w Wd D pLSA provides a generative model only for the training data LDA treats P(zk |di ) as a latent random variable, sampling it from a Dirichlet distribution P(θ|α) Maximizes data likelihood P(w|φ, α, β) = P(w|z, φ)P(z|θ)P(θ|α)P(φ|β)dθ z Davide Bacciu Probabilistic Generative Models for Machine Vision
- 17. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Hierarchical Latent Topic Models... Dependent Pachinko Allocation Model (Li and McCallum, 2006) Davide Bacciu Probabilistic Generative Models for Machine Vision
- 18. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Image Annotation Learn association between image content and text Typically exploits region-based image representation Image segmentation Color and texture descriptors Non-probabilistic image annotation Information retrieval approaches (e.g. vector based) Multiple instance learning Probabilistic Image Annotation Machine translation approach Probabilistic generative Latent topic models Davide Bacciu Probabilistic Generative Models for Machine Vision
- 19. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Latent Aspect Models in Image Annotation µ µ z b α θ z b Bd σ Bd σ α θ v t β v t β Md Md D D GM-LDA CORR-LDA Based on a Mixture of Gaussians modeling of the Bd image regions and a multinomial distribution for annotations Md Davide Bacciu Probabilistic Generative Models for Machine Vision
- 20. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications Image Annotation - Is it Worth the Hassle? The Web is full of annotated multimedia information Davide Bacciu Probabilistic Generative Models for Machine Vision
- 21. Introduction Machine Vision Background Visual Content Representation Region Annotation by Hierarchical Region Topic Discovery Machine Vision Applications How Informative is Image Representation? BOW assumption is strong in the visual domain: spatial context matters! Region-based representation is often too coarse and unstable. How can we retain the robustness of BOW while introducing spatial context? Davide Bacciu Probabilistic Generative Models for Machine Vision
- 22. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Introduction Overtake flat region-based and keypoint-based approaches Structured representation conveying richer semantics and discriminative power than a BOW model without the rigidity of a part-based approach Key idea Introduce a multi-resolution representation of visual content drawing both on region-based and keypoint-based models Introduce a multi-layered structure at the semantic level of latent topic discovery Davide Bacciu Probabilistic Generative Models for Machine Vision
- 23. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Multi-Resolution Image Representation Multi-resolution describing visual information at different spatial granularity BOW representation is based on assumption that spatial distribution of visual keywords can be neglected when determining the overall image appearance Biased view holding mostly for corpora characterized by small variability of the pictorial content Need of structured approach that can isolate semantically different image portions Text representation Documents are collections of paragraphs Paragraphs are collections of words Davide Bacciu Probabilistic Generative Models for Machine Vision
- 24. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Region-Keypoints Representation r1 r2 r3 r4 wr bag−of−words r5 segmented image ... bag−of−regions Use an unsupervised learning model to extract perceptually coherent portions of the image rl Use dense sampling to extract a set of local visual descriptors Davide Bacciu Probabilistic Generative Models for Machine Vision
- 25. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Hierarchical Region-Topic Probabilistic Latent Semantic Analysis (HRPLSA) φ T d z t w Wr rd D Complete model likelihood nilj D Ri K Wl T LC (N |θ) = P(zk |di )Zilk P(tp |zk )(φjp )Tkjp i=1 l=1 k =1 j=1 p=1 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 26. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Model Fitting HRPLSA parameters are fitted by applying Expectation Maximization to the complete log-likelihood, e.g. D Ri Wl i=1 l=1 P(zk |di , rl ) j=1 nilj P(tp |zk , wj ) P(tp |zk ) = Ri D i =1 l =1 P(zk |di , rl )ni l · Model parameters θ P(zk |di ) - Document-specific mixing proportions of region topics P(tp |zk ) - Mixing proportions of keyword topics given region aspects φjp - Multinomial keyword-topic distribution Probabilistic inference is extended to images outside the training pool by folding-in Davide Bacciu Probabilistic Generative Models for Machine Vision
- 27. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Modeling Images and Words by HRPLSA (HRPLSA-ANN) d sky building P (a|z) streetlight sing z1 z2 ... zR awning ... window P (a|t) sidewalk t1 t2 ... tW1 t1 t2 ... tWR plants car road w1 w2 wW1 w1 w2 wWR Connection between region-topics (keyword-topics) and word is learned by maximizing constrained annotation log-likelihood D F K Lcal (N |θ ) = mfi log P(af |zk )P(zk |di ) i=1 f =1 k =1 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 28. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Probabilistic Inference A Few Examples Most likely visual aspect within an image di z i = arg max P(zk |di ) zk ∈Z Most likely visual aspect z ∗ for an image segment rl z ∗ = arg max P(zk |di , rl ) zk ∈Z Most likely annotation a∗ for region rl in image dnew K a∗ = arg max P(af |zk )P(zk |dnew ) af ∈A k =1 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 29. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Unsupervised Discovery and Segmentation of Visual Classes Microsoft Research Cambridge dataset (MSRC-B1) 240 images and 9 visual classes Ground truth segmentation of visual classes Image segments are assigned to the most likely discovered aspect and segmentation overlap is measured GTo ∩ Rr ρor = . GTo ∪ Rr Davide Bacciu Probabilistic Generative Models for Machine Vision
- 30. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Semantic Segmentation Discovery of semantic visual aspects corrects segmentation errors Davide Bacciu Probabilistic Generative Models for Machine Vision
- 31. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation MSRC-B1 Semantic Segmentation Result Table: Segmentation Overlap on the MSRC-B1 Dataset Algorithm Airplanes Bicycles Buildings Cars Cows Faces Grass Tree Sky Average dLDA 0.43 0.50 0.16 0.45 0.52 0.44 0.60 0.71 0.74 0.51 LDA10 0.11 0.06 0.09 0.14 0.11 0.40 0.41 0.69 0.41 0.27 LDA15 0.08 0.56 0.06 0.14 0.14 0.43 0.57 0.62 0.41 0.33 LDA20 0.10 0.50 0.40 0.15 0.58 0.45 0.54 0.46 0.50 0.40 LDA25 0.14 0.52 0.21 0.17 0.48 0.44 0.45 0.59 0.50 0.39 HPRLSA10 15 0.32 0.37 0.59 0.68 0.74 0.44 0.87 0.81 0.90 0.62 HPRLSA10 20 0.35 0.56 0.65 0.67 0.74 0.47 0.89 0.47 0.91 0.63 HPRLSA10 25 0.35 0.52 0.75 0.63 0.73 0.50 0.89 0.71 0.86 0.66 HPRLSA15 20 0.36 0.49 0.56 0.60 0.68 0.52 0.89 0.77 0.91 0.64 HPRLSA15 25 0.42 0.52 0.61 0.68 0.69 0.45 0.90 0.61 0.92 0.65 HPRLSA20 25 0.44 0.43 0.48 0.52 0.67 0.48 0.88 0.75 0.91 0.61 10 HPRLSA25 0.16 0.32 0.15 0.40 0.43 0.22 0.28 0.33 0.38 0.30 15 HPRLSA25 0.24 0.31 0.13 0.38 0.37 0.19 0.25 0.21 0.45 0.28 Comparative analysis HRPLSA Multi-segmentation Latent Dirichlet Allocation (LDA) Hierarchical Latent Dirichlet Allocation (hLDA) Davide Bacciu Probabilistic Generative Models for Machine Vision
- 32. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation MSRC-B1 Segmentation Example I Davide Bacciu Probabilistic Generative Models for Machine Vision
- 33. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation MSRC-B1 Segmentation Example II Davide Bacciu Probabilistic Generative Models for Machine Vision
- 34. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation MSRC-B1 Segmentation Example III Davide Bacciu Probabilistic Generative Models for Machine Vision
- 35. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Image Annotation Washington Ground Truth dataset 854 images manually annotated with 392 words 427 training/test images LabelMe dataset 2688 images annotated with ∼ 400 words 800 training / 1888 test images Comparative analysis HRPLSA-ANN pLSA-FEATURES Annotation Propagation (AP) Precision/recall analysis Davide Bacciu Probabilistic Generative Models for Machine Vision
- 36. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Washington Dataset Image Annotation Precision/Recall Plot Comparative precision/recall results on the Washington dataset 0.8 plsa245 0.75 hrplsa245 ap245 0.7 plsa188 0.65 hrplsa188 ap188 0.6 precision plsa158 hrplsa158 0.55 ap158 0.5 0.45 0.4 0.35 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 recall Davide Bacciu Probabilistic Generative Models for Machine Vision
- 37. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Washington Dataset Image Annotation Example Davide Bacciu Probabilistic Generative Models for Machine Vision
- 38. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Washington Dataset Region Annotation Example (I) Examples of good HRPLSA-ANN annotations 50 50 clouds tree clouds tree 100 100 building 150 tree tree 150 tree 200 clouds 200 clouds sky 250 tree 250 tree 300 300 house 350 350 tree 400 400 people tree sidewalk tree people 450 450 500 50 clear 100 clear house 150 window 200 250 tree building tree 300 building 350 tree 400 tree tree tree 450 building tree 500 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 39. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Washington Dataset Region Annotation Example (II) Examples of bad HRPLSA-ANN annotations 50 50 tree clouds tree tree clouds clouds 100 100 tree 150 tree 150 200 200 water sky tree tree 250 250 tree tree tree tree 300 300 tree 350 tree 350 sky 400 400 tree sky tree bush 450 450 flower 500 500 50 overcast sky people 100 150 house clouds 200 250 tree people 300 350 tree 400 tree tree 450 500 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 40. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation Washington Dataset Annotated Segments Davide Bacciu Probabilistic Generative Models for Machine Vision
- 41. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation LabelMe Dataset Image Annotation Precision/Recall Plot 0.8 plsa60 0.75 hrplsa60,80 plsa25 0.7 hrplsa25,50 ap 0.65 Precision 0.6 0.55 0.5 0.45 0.4 0.35 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Recall Davide Bacciu Probabilistic Generative Models for Machine Vision
- 42. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation LabelMe Dataset Example of Annotated Images Davide Bacciu Probabilistic Generative Models for Machine Vision
- 43. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation LabelMe Dataset Recalling Specific Terms Balcony Sign balcony sign 0.45 0.45 hrplsa hrplsa 0.4 plsa 0.4 plsa ap ap 0.35 0.35 0.3 0.3 precision/recall precision/recall 0.25 0.25 0.2 0.2 0.15 0.15 0.1 0.1 0.05 0.05 0 0 precision recall precision recall Davide Bacciu Probabilistic Generative Models for Machine Vision
- 44. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation LabelMe Dataset Region Annotation Issue I Region annotation tends to predict only a limited number of words 0.16 coast forest 400 2000 0.14 200 1000 0.12 0 0 0 100 200 300 0 100 200 300 highway city 0.1 1000 2000 500 1000 0.08 0 0 0 100 200 300 0 100 200 300 0.06 mountain country 1000 1000 0.04 500 500 0.02 0 0 0 100 200 300 0 100 200 300 street tallbuilding 0 2000 2000 g n on er ld ad w e er r a y ca sk in ai tre se do fie ap at rs ro ild nt 1000 1000 w in pe cr ou bu w ys m sk 0 0 0 100 200 300 0 100 200 300 Davide Bacciu Probabilistic Generative Models for Machine Vision
- 45. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation LabelMe Dataset Region Annotation Issue II Distribution of region-topics in mountain caption Topic 11 Topic 40 Topic 20 The unbalanced distribution of term occurrences introduces a bias in the generation of the segment caption that yields to distinct region aspects being associated to the same frequently co-occurring word Davide Bacciu Probabilistic Generative Models for Machine Vision
- 46. Visual Content Representation Machine Vision Background Hierarchical Probabilistic Latent Semantic Analysis Region Annotation by Hierarchical Region Topic Discovery Experimental Evaluation LabelMe Dataset Annotated Segments Davide Bacciu Probabilistic Generative Models for Machine Vision
- 47. Concluding Remarks Conclusion Bibliography and Online Resources Wrapping up... Inspiration can be sought in related application fields Bag of words representation Generative models for text Need to pay attention to the underlying assumptions of a model! Hierarchical multi-resolution latent aspect model (HRPLSA) Unsupervised semantic segmentation of visual content Multi-level image captioning Future challenges Generalize multi-resolution representation and topic hierarchy beyond a two-layered structure Video/multimedia understanding Davide Bacciu Probabilistic Generative Models for Machine Vision
- 48. Concluding Remarks Conclusion Bibliography and Online Resources Online Resources A notable introductory course to object recognition (with code) http://people.csail.mit.edu/torralba/ shortCourseRLOC/ Code repositories http://www.cs.ubc.ca/~lowe/keypoints/ (Original SIFT software by D. Lowe) http://vlfeat.org/ (SIFT, MSER and VQ routines) www.robots.ox.ac.uk/~vgg/research/affine/ (Affine detectors and more) http://www.di.ens.fr/~russell/projects/mult_ seg_discovery/index.html (Multiple segmentation LDA) Image annotation tool and repository http://labelme.csail.mit.edu/ (LabelMe) Davide Bacciu Probabilistic Generative Models for Machine Vision
- 49. Concluding Remarks Conclusion Bibliography and Online Resources Bibliography D. Bacciu, A Perceptual Learning Model to Discover the Hierarchical Latent Structure of Image Collections, Ph.D. Thesis, 2008. J. Sivic and A. Zisserman. Video google: a text retrieval approach to object matching in videos. Proceedings of the Ninth IEEE International Conference on Computer Vision, ICCV 2003, pages 1470–1477, vol.2, Oct. 2003. J. Sivic, B.C. Russell, A.A. Efros, A. Zisserman, and W.T. Freeman. Discovering objects and their location in images. Proceedings of the Tenth IEEE International Conference on Computer Vision, ICCV 2005, 370–377, Oct. 2005. J. Sivic, B. C. Russell, A. Zisserman, W. T. Freeman, and A. A. Efros. Unsupervised discovery of visual object class hierarchies. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2008. K. Barnard, P. Duygulu, N. de Freitas, D. Forsyth, D. M. Blei, and M. I. Jordan. Matching words and pictures. Journal of Machine Learning Research, vol. 3, 1107–1135, Feb. 2003. D. M. Blei and M. I. Jordan. Modeling annotated data. Proceedings of the International Conference on Research and Development in Information Retrieval, SIGIR 2003, Aug. 2003. F. Monay and D. Gatica-Perez. Modeling semantic aspects for cross-media image indexing. IEEE Transactions on Pattern Analysis and Machine Intelligence, 29(10):1802–1817, Oct. 2007. Davide Bacciu Probabilistic Generative Models for Machine Vision
- 50. Concluding Remarks Conclusion Bibliography and Online Resources Thank You Davide Bacciu Probabilistic Generative Models for Machine Vision