VConv-DAE: Deep Volumetric Shape Learning Without Object Labels View Full Text


Ontology type: schema:Chapter      Open Access: True


Chapter Info

DATE

2016-11-24

AUTHORS

Abhishek Sharma , Oliver Grau , Mario Fritz

ABSTRACT

With the advent of affordable depth sensors, 3D capture becomes more and more ubiquitous and already has made its way into commercial products. Yet, capturing the geometry or complete shapes of everyday objects using scanning devices (e.g. Kinect) still comes with several challenges that result in noise or even incomplete shapes.Recent success in deep learning has shown how to learn complex shape distributions in a data-driven way from large scale 3D CAD Model collections and to utilize them for 3D processing on volumetric representations and thereby circumventing problems of topology and tessellation. Prior work has shown encouraging results on problems ranging from shape completion to recognition. We provide an analysis of such approaches and discover that training as well as the resulting representation are strongly and unnecessarily tied to the notion of object labels. Thus, we propose a full convolutional volumetric auto encoder that learns volumetric representation from noisy data by estimating the voxel occupancy grids. The proposed method outperforms prior work on challenging tasks like denoising and shape completion. We also show that the obtained deep embedding gives competitive performance when used for classification and promising results for shape interpolation. More... »

PAGES

236-250

Identifiers

URI

http://scigraph.springernature.com/pub.10.1007/978-3-319-49409-8_20

DOI

http://dx.doi.org/10.1007/978-3-319-49409-8_20

DIMENSIONS

https://app.dimensions.ai/details/publication/pub.1084689861


Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
Incoming Citations Browse incoming citations for this publication using opencitations.net

JSON-LD is the canonical representation for SciGraph data.

TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

[
  {
    "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
    "about": [
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Information and Computing Sciences", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Artificial Intelligence and Image Processing", 
        "type": "DefinedTerm"
      }
    ], 
    "author": [
      {
        "affiliation": {
          "alternateName": "Intel Visual Computing Institute, Saarbr\u00fccken, Germany", 
          "id": "http://www.grid.ac/institutes/None", 
          "name": [
            "Intel Visual Computing Institute, Saarbr\u00fccken, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Sharma", 
        "givenName": "Abhishek", 
        "id": "sg:person.07441524645.18", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.07441524645.18"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Intel, Saarbr\u00fccken, Germany", 
          "id": "http://www.grid.ac/institutes/grid.425153.4", 
          "name": [
            "Intel, Saarbr\u00fccken, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Grau", 
        "givenName": "Oliver", 
        "id": "sg:person.07536663710.54", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.07536663710.54"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Max Planck Institute for Informatics, Saarbr\u00fccken, Germany", 
          "id": "http://www.grid.ac/institutes/grid.419528.3", 
          "name": [
            "Max Planck Institute for Informatics, Saarbr\u00fccken, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Fritz", 
        "givenName": "Mario", 
        "id": "sg:person.013361072755.17", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013361072755.17"
        ], 
        "type": "Person"
      }
    ], 
    "datePublished": "2016-11-24", 
    "datePublishedReg": "2016-11-24", 
    "description": "With the advent of affordable depth sensors, 3D capture becomes more and more ubiquitous and already has made its way into commercial products. Yet, capturing the geometry or complete shapes of everyday objects using scanning devices (e.g. Kinect) still comes with several challenges that result in noise or even incomplete shapes.Recent success in deep learning has shown how to learn complex shape distributions in a data-driven way from large scale 3D CAD Model collections and to utilize them for 3D processing on volumetric representations and thereby circumventing problems of topology and tessellation. Prior work has shown encouraging results on problems ranging from shape completion to recognition. We provide an analysis of such approaches and discover that training as well as the resulting representation are strongly and unnecessarily tied to the notion of object labels. Thus, we propose a full convolutional volumetric auto encoder that learns volumetric representation from noisy data by estimating the voxel occupancy grids. The proposed method outperforms prior work on challenging tasks like denoising and shape completion. We also show that the obtained deep embedding gives competitive performance when used for classification and promising results for shape interpolation.", 
    "editor": [
      {
        "familyName": "Hua", 
        "givenName": "Gang", 
        "type": "Person"
      }, 
      {
        "familyName": "J\u00e9gou", 
        "givenName": "Herv\u00e9", 
        "type": "Person"
      }
    ], 
    "genre": "chapter", 
    "id": "sg:pub.10.1007/978-3-319-49409-8_20", 
    "isAccessibleForFree": true, 
    "isPartOf": {
      "isbn": [
        "978-3-319-49408-1", 
        "978-3-319-49409-8"
      ], 
      "name": "Computer Vision \u2013 ECCV 2016 Workshops", 
      "type": "Book"
    }, 
    "keywords": [
      "volumetric representation", 
      "shape completion", 
      "object labels", 
      "affordable depth sensors", 
      "voxel occupancy grid", 
      "prior work", 
      "data-driven way", 
      "deep learning", 
      "occupancy grid", 
      "model collections", 
      "Auto-Encoder", 
      "deep embedding", 
      "depth sensor", 
      "challenging task", 
      "competitive performance", 
      "everyday objects", 
      "noisy data", 
      "incomplete shapes", 
      "volumetric shape", 
      "shape interpolation", 
      "complete shape", 
      "problems of topology", 
      "such approaches", 
      "recent success", 
      "promising results", 
      "scanning device", 
      "representation", 
      "encouraging results", 
      "labels", 
      "encoder", 
      "embedding", 
      "task", 
      "learning", 
      "objects", 
      "classification", 
      "recognition", 
      "commercial products", 
      "grid", 
      "processing", 
      "topology", 
      "sensors", 
      "way", 
      "work", 
      "advent", 
      "interpolation", 
      "performance", 
      "tessellation", 
      "challenges", 
      "collection", 
      "devices", 
      "noise", 
      "training", 
      "shape distribution", 
      "data", 
      "results", 
      "method", 
      "notion", 
      "capture", 
      "shape", 
      "success", 
      "completion", 
      "analysis", 
      "products", 
      "geometry", 
      "distribution", 
      "problem", 
      "approach"
    ], 
    "name": "VConv-DAE: Deep Volumetric Shape Learning Without Object Labels", 
    "pagination": "236-250", 
    "productId": [
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "pub.1084689861"
        ]
      }, 
      {
        "name": "doi", 
        "type": "PropertyValue", 
        "value": [
          "10.1007/978-3-319-49409-8_20"
        ]
      }
    ], 
    "publisher": {
      "name": "Springer Nature", 
      "type": "Organisation"
    }, 
    "sameAs": [
      "https://doi.org/10.1007/978-3-319-49409-8_20", 
      "https://app.dimensions.ai/details/publication/pub.1084689861"
    ], 
    "sdDataset": "chapters", 
    "sdDatePublished": "2022-12-01T06:53", 
    "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
    "sdPublisher": {
      "name": "Springer Nature - SN SciGraph project", 
      "type": "Organization"
    }, 
    "sdSource": "s3://com-springernature-scigraph/baseset/20221201/entities/gbq_results/chapter/chapter_428.jsonl", 
    "type": "Chapter", 
    "url": "https://doi.org/10.1007/978-3-319-49409-8_20"
  }
]
 

Download the RDF metadata as:  json-ld nt turtle xml License info

HOW TO GET THIS DATA PROGRAMMATICALLY:

JSON-LD is a popular format for linked data which is fully compatible with JSON.

curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-49409-8_20'

N-Triples is a line-based linked data format ideal for batch operations.

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-49409-8_20'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-49409-8_20'

RDF/XML is a standard XML format for linked data.

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-49409-8_20'


 

This table displays all metadata directly associated to this object as RDF triples.

151 TRIPLES      22 PREDICATES      91 URIs      84 LITERALS      7 BLANK NODES

Subject Predicate Object
1 sg:pub.10.1007/978-3-319-49409-8_20 schema:about anzsrc-for:08
2 anzsrc-for:0801
3 schema:author N388ee48df2874edda47c3ed12f8b31e2
4 schema:datePublished 2016-11-24
5 schema:datePublishedReg 2016-11-24
6 schema:description With the advent of affordable depth sensors, 3D capture becomes more and more ubiquitous and already has made its way into commercial products. Yet, capturing the geometry or complete shapes of everyday objects using scanning devices (e.g. Kinect) still comes with several challenges that result in noise or even incomplete shapes.Recent success in deep learning has shown how to learn complex shape distributions in a data-driven way from large scale 3D CAD Model collections and to utilize them for 3D processing on volumetric representations and thereby circumventing problems of topology and tessellation. Prior work has shown encouraging results on problems ranging from shape completion to recognition. We provide an analysis of such approaches and discover that training as well as the resulting representation are strongly and unnecessarily tied to the notion of object labels. Thus, we propose a full convolutional volumetric auto encoder that learns volumetric representation from noisy data by estimating the voxel occupancy grids. The proposed method outperforms prior work on challenging tasks like denoising and shape completion. We also show that the obtained deep embedding gives competitive performance when used for classification and promising results for shape interpolation.
7 schema:editor Nb760d7479b264b599dc70b051b22ae71
8 schema:genre chapter
9 schema:isAccessibleForFree true
10 schema:isPartOf N99c701f1783c48cdb56df34cbfd702e5
11 schema:keywords Auto-Encoder
12 advent
13 affordable depth sensors
14 analysis
15 approach
16 capture
17 challenges
18 challenging task
19 classification
20 collection
21 commercial products
22 competitive performance
23 complete shape
24 completion
25 data
26 data-driven way
27 deep embedding
28 deep learning
29 depth sensor
30 devices
31 distribution
32 embedding
33 encoder
34 encouraging results
35 everyday objects
36 geometry
37 grid
38 incomplete shapes
39 interpolation
40 labels
41 learning
42 method
43 model collections
44 noise
45 noisy data
46 notion
47 object labels
48 objects
49 occupancy grid
50 performance
51 prior work
52 problem
53 problems of topology
54 processing
55 products
56 promising results
57 recent success
58 recognition
59 representation
60 results
61 scanning device
62 sensors
63 shape
64 shape completion
65 shape distribution
66 shape interpolation
67 success
68 such approaches
69 task
70 tessellation
71 topology
72 training
73 volumetric representation
74 volumetric shape
75 voxel occupancy grid
76 way
77 work
78 schema:name VConv-DAE: Deep Volumetric Shape Learning Without Object Labels
79 schema:pagination 236-250
80 schema:productId Nc3858f12ad2f4c5f9fe82efb71c8b175
81 Nf50d27193f044e7fba64e162aa5a9918
82 schema:publisher Nc49fcd6d974f439b87ef0f390f828e47
83 schema:sameAs https://app.dimensions.ai/details/publication/pub.1084689861
84 https://doi.org/10.1007/978-3-319-49409-8_20
85 schema:sdDatePublished 2022-12-01T06:53
86 schema:sdLicense https://scigraph.springernature.com/explorer/license/
87 schema:sdPublisher N2a87951dad7a49928685f67df75e8a9f
88 schema:url https://doi.org/10.1007/978-3-319-49409-8_20
89 sgo:license sg:explorer/license/
90 sgo:sdDataset chapters
91 rdf:type schema:Chapter
92 N25f22add7aac48aeb6468041a2558756 schema:familyName Jégou
93 schema:givenName Hervé
94 rdf:type schema:Person
95 N2a87951dad7a49928685f67df75e8a9f schema:name Springer Nature - SN SciGraph project
96 rdf:type schema:Organization
97 N388ee48df2874edda47c3ed12f8b31e2 rdf:first sg:person.07441524645.18
98 rdf:rest N3cb64e59e59c4ca4be148e3a00df852d
99 N3cb64e59e59c4ca4be148e3a00df852d rdf:first sg:person.07536663710.54
100 rdf:rest Nd9b78a3c8a8d40b3a0c21c652002c39c
101 N99c701f1783c48cdb56df34cbfd702e5 schema:isbn 978-3-319-49408-1
102 978-3-319-49409-8
103 schema:name Computer Vision – ECCV 2016 Workshops
104 rdf:type schema:Book
105 Na13577b5b42b471aaa20cf9ee83d0cfc rdf:first N25f22add7aac48aeb6468041a2558756
106 rdf:rest rdf:nil
107 Nb760d7479b264b599dc70b051b22ae71 rdf:first Neb75accb1cd6470cbce15606f6a57a54
108 rdf:rest Na13577b5b42b471aaa20cf9ee83d0cfc
109 Nc3858f12ad2f4c5f9fe82efb71c8b175 schema:name dimensions_id
110 schema:value pub.1084689861
111 rdf:type schema:PropertyValue
112 Nc49fcd6d974f439b87ef0f390f828e47 schema:name Springer Nature
113 rdf:type schema:Organisation
114 Nd9b78a3c8a8d40b3a0c21c652002c39c rdf:first sg:person.013361072755.17
115 rdf:rest rdf:nil
116 Neb75accb1cd6470cbce15606f6a57a54 schema:familyName Hua
117 schema:givenName Gang
118 rdf:type schema:Person
119 Nf50d27193f044e7fba64e162aa5a9918 schema:name doi
120 schema:value 10.1007/978-3-319-49409-8_20
121 rdf:type schema:PropertyValue
122 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
123 schema:name Information and Computing Sciences
124 rdf:type schema:DefinedTerm
125 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
126 schema:name Artificial Intelligence and Image Processing
127 rdf:type schema:DefinedTerm
128 sg:person.013361072755.17 schema:affiliation grid-institutes:grid.419528.3
129 schema:familyName Fritz
130 schema:givenName Mario
131 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013361072755.17
132 rdf:type schema:Person
133 sg:person.07441524645.18 schema:affiliation grid-institutes:None
134 schema:familyName Sharma
135 schema:givenName Abhishek
136 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.07441524645.18
137 rdf:type schema:Person
138 sg:person.07536663710.54 schema:affiliation grid-institutes:grid.425153.4
139 schema:familyName Grau
140 schema:givenName Oliver
141 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.07536663710.54
142 rdf:type schema:Person
143 grid-institutes:None schema:alternateName Intel Visual Computing Institute, Saarbrücken, Germany
144 schema:name Intel Visual Computing Institute, Saarbrücken, Germany
145 rdf:type schema:Organization
146 grid-institutes:grid.419528.3 schema:alternateName Max Planck Institute for Informatics, Saarbrücken, Germany
147 schema:name Max Planck Institute for Informatics, Saarbrücken, Germany
148 rdf:type schema:Organization
149 grid-institutes:grid.425153.4 schema:alternateName Intel, Saarbrücken, Germany
150 schema:name Intel, Saarbrücken, Germany
151 rdf:type schema:Organization
 




Preview window. Press ESC to close (or click here)


...