%PDF-1.3 1 0 obj << /Kids [ 4 0 R 5 0 R 6 0 R 7 0 R 8 0 R 9 0 R 10 0 R 11 0 R 12 0 R ] /Type /Pages /Count 9 >> endobj 2 0 obj << /Subject (Neural Information Processing Systems http\072\057\057nips\056cc\057) /Publisher (Curran Associates) /Language (en\055US) /Created (2011) /Description-Abstract (We present a joint image segmentation and labeling model \050JSL\051 which\054 given a bag of figure\055ground segment hypotheses extracted at multiple image locations and scales\054 constructs a joint probability distribution over both the compatible image interpretations \050tilings or image segmentations\051 composed from those segments\054 and over their labeling into categories\056 The process of drawing samples from the joint distribution can be interpreted as first sampling tilings\054 modeled as maximal cliques\054 from a graph connecting spatially non\055overlapping segments in the bag\054 followed by sampling labels for those segments\054 conditioned on the choice of a particular tiling\056 We learn the segmentation and labeling parameters jointly\054 based on Maximum Likelihood with a novel Incremental Saddle Point estimation procedure\056 The partition function over tilings and labelings is increasingly more accurately approximated by including incorrect configurations that a not\055yet\055competent model rates probable during learning\056 We show that the proposed methodology matches the current state of the art in the Stanford dataset\054 as well as in VOC2010\054 where 41\0567\045 accuracy on the test set is achieved\056) /Producer (Python PDF Library \055 http\072\057\057pybrary\056net\057pyPdf\057) /Title (Probabilistic Joint Image Segmentation and Labeling) /Date (2011) /Type (Conference Proceedings) /firstpage (1827) /Book (Advances in Neural Information Processing Systems 24) /Description (Paper accepted and presented at the Neural Information Processing Systems Conference \050http\072\057\057nips\056cc\057\051) /Editors (J\056 Shawe\055Taylor and R\056S\056 Zemel and P\056L\056 Bartlett and F\056 Pereira and K\056Q\056 Weinberger) /Author (Adrian Ion\054 Joao Carreira\054 Cristian Sminchisescu) /lastpage (1835) >> endobj 3 0 obj << /Type /Catalog /Pages 1 0 R >> endobj 4 0 obj << /Parent 1 0 R /Rotate 0 /Contents 13 0 R /Resources << /ExtGState 14 0 R /ProcSet [ /PDF /ImageB /Text ] /Font 16 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 5 0 obj << /Parent 1 0 R /Rotate 0 /Contents 44 0 R /Resources << /ExtGState 45 0 R /XObject 47 0 R /ProcSet [ /PDF /ImageC /Text ] /Font 68 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 6 0 obj << /Parent 1 0 R /Rotate 0 /Contents 90 0 R /Resources << /ExtGState 91 0 R /ProcSet [ /PDF /ImageB /Text ] /Font 92 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 7 0 obj << /Parent 1 0 R /Rotate 0 /Contents 107 0 R /Resources << /ExtGState 108 0 R /ProcSet [ /PDF /Text ] /Font 109 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 8 0 obj << /Parent 1 0 R /Rotate 0 /Contents 110 0 R /Resources << /ExtGState 111 0 R /ProcSet [ /PDF /ImageB /Text ] /Font 112 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 9 0 obj << /Parent 1 0 R /Rotate 0 /Contents 116 0 R /Resources << /ExtGState 117 0 R /ProcSet [ /PDF /ImageB /Text ] /Font 118 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 10 0 obj << /Parent 1 0 R /Rotate 0 /Contents 119 0 R /Resources << /ExtGState 120 0 R /XObject 121 0 R /ProcSet [ /PDF /ImageB /ImageC /Text ] /Font 128 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 11 0 obj << /Parent 1 0 R /Rotate 0 /Contents 131 0 R /Resources << /ExtGState 132 0 R /ProcSet [ /PDF /Text ] /Font 133 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 12 0 obj << /Parent 1 0 R /Rotate 0 /Contents 134 0 R /Resources << /ExtGState 135 0 R /ProcSet [ /PDF /Text ] /Font 136 0 R >> /MediaBox [ 0 0 612 792 ] /Type /Page >> endobj 13 0 obj << /Length 7467 /Filter /FlateDecode >> stream x=ْƑm_v{C Nv_v)aje[F~<,rHQ-of,{8Z#H ԑv v#^xiw'v^|SΆ8;1Nn4q3ţ'ٽ7;q?r/~]1N{| s ѻg">.8蝵nrw/G9Q*w09oG1̓1+j)wκ|naa{4Rw +PQ<9شx>z&vjƱa4gzpT6,J״YûJ,