Shadow-Catcher: Looking into Shadows to Detect Ghost Objects in Autonomous Vehicle 3D Sensing View Full Text


Ontology type: schema:Chapter      Open Access: True


Chapter Info

DATE

2021-09-30

AUTHORS

Zhongyuan Hau , Soteris Demetriou , Luis Muñoz-González , Emil C. Lupu

ABSTRACT

LiDAR-driven 3D sensing allows new generations of vehicles to achieve advanced levels of situation awareness. However, recent works have demonstrated that physical adversaries can spoof LiDAR return signals and deceive 3D object detectors to erroneously detect “ghost" objects. Existing defenses are either impractical or focus only on vehicles. Unfortunately, it is easier to spoof smaller objects such as pedestrians and cyclists, but harder to defend against and can have worse safety implications. To address this gap, we introduce Shadow-Catcher, a set of new techniques embodied in an end-to-end prototype to detect both large and small ghost object attacks on 3D detectors. We characterize a new semantically meaningful physical invariant (3D shadows) which Shadow-Catcher leverages for validating objects. Our evaluation on the KITTI dataset shows that Shadow-Catcher consistently achieves more than 94% accuracy in identifying anomalous shadows for vehicles, pedestrians, and cyclists, while it remains robust to a novel class of strong “invalidation” attacks targeting the defense system. Shadow-Catcher can achieve real-time detection, requiring only between 0.003 s–0.021 s on average to process an object in a 3D point cloud on commodity hardware and achieves a 2.17x speedup compared to prior work. More... »

PAGES

691-711

Book

TITLE

Computer Security – ESORICS 2021

ISBN

978-3-030-88417-8
978-3-030-88418-5

Identifiers

URI

http://scigraph.springernature.com/pub.10.1007/978-3-030-88418-5_33

DOI

http://dx.doi.org/10.1007/978-3-030-88418-5_33

DIMENSIONS

https://app.dimensions.ai/details/publication/pub.1141490051


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": "Imperial College London, London, UK", 
          "id": "http://www.grid.ac/institutes/grid.7445.2", 
          "name": [
            "Imperial College London, London, UK"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Hau", 
        "givenName": "Zhongyuan", 
        "id": "sg:person.014731402615.26", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014731402615.26"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Imperial College London, London, UK", 
          "id": "http://www.grid.ac/institutes/grid.7445.2", 
          "name": [
            "Imperial College London, London, UK"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Demetriou", 
        "givenName": "Soteris", 
        "id": "sg:person.011365416202.34", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011365416202.34"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Imperial College London, London, UK", 
          "id": "http://www.grid.ac/institutes/grid.7445.2", 
          "name": [
            "Imperial College London, London, UK"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Mu\u00f1oz-Gonz\u00e1lez", 
        "givenName": "Luis", 
        "id": "sg:person.011755172717.81", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011755172717.81"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Imperial College London, London, UK", 
          "id": "http://www.grid.ac/institutes/grid.7445.2", 
          "name": [
            "Imperial College London, London, UK"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Lupu", 
        "givenName": "Emil C.", 
        "id": "sg:person.013404167044.28", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013404167044.28"
        ], 
        "type": "Person"
      }
    ], 
    "datePublished": "2021-09-30", 
    "datePublishedReg": "2021-09-30", 
    "description": "LiDAR-driven 3D sensing allows new generations of vehicles to achieve advanced levels of situation awareness. However, recent works have demonstrated that physical adversaries can spoof LiDAR return signals and deceive 3D object detectors to erroneously detect \u201cghost\" objects. Existing defenses are either impractical or focus only on vehicles. Unfortunately, it is easier to spoof smaller objects such as pedestrians and cyclists, but harder to defend against and can have worse safety implications. To address this gap, we introduce Shadow-Catcher, a set of new techniques embodied in an end-to-end prototype to detect both large and small ghost object attacks on 3D detectors. We characterize a new semantically meaningful physical invariant (3D shadows) which Shadow-Catcher leverages for validating objects. Our evaluation on the KITTI dataset shows that Shadow-Catcher consistently achieves more than 94% accuracy in identifying anomalous shadows for vehicles, pedestrians, and cyclists, while it remains robust to a novel class of strong \u201cinvalidation\u201d attacks targeting the defense system. Shadow-Catcher can achieve real-time detection, requiring only between 0.003\u00a0s\u20130.021\u00a0s on average to process an object in a 3D point cloud on commodity hardware and achieves a 2.17x speedup compared to prior work.", 
    "editor": [
      {
        "familyName": "Bertino", 
        "givenName": "Elisa", 
        "type": "Person"
      }, 
      {
        "familyName": "Shulman", 
        "givenName": "Haya", 
        "type": "Person"
      }, 
      {
        "familyName": "Waidner", 
        "givenName": "Michael", 
        "type": "Person"
      }
    ], 
    "genre": "chapter", 
    "id": "sg:pub.10.1007/978-3-030-88418-5_33", 
    "isAccessibleForFree": true, 
    "isPartOf": {
      "isbn": [
        "978-3-030-88417-8", 
        "978-3-030-88418-5"
      ], 
      "name": "Computer Security \u2013 ESORICS 2021", 
      "type": "Book"
    }, 
    "keywords": [
      "commodity hardware", 
      "KITTI dataset", 
      "physical adversaries", 
      "object detector", 
      "real-time detection", 
      "situation awareness", 
      "ghost objects", 
      "point clouds", 
      "end prototype", 
      "prior work", 
      "small objects", 
      "objects", 
      "attacks", 
      "pedestrians", 
      "vehicles", 
      "speedup", 
      "adversary", 
      "hardware", 
      "new generation", 
      "dataset", 
      "cloud", 
      "physical invariants", 
      "new technique", 
      "recent work", 
      "prototype", 
      "accuracy", 
      "shadow", 
      "advanced level", 
      "set", 
      "work", 
      "invalidation", 
      "lidar", 
      "sensing", 
      "safety implications", 
      "leverage", 
      "detection", 
      "system", 
      "detector", 
      "technique", 
      "return signal", 
      "invariants", 
      "defense system", 
      "generation", 
      "class", 
      "evaluation", 
      "signals", 
      "end", 
      "awareness", 
      "gap", 
      "defense", 
      "cyclists", 
      "lidar return signals", 
      "novel class", 
      "ghosts", 
      "levels", 
      "implications"
    ], 
    "name": "Shadow-Catcher: Looking into Shadows to Detect Ghost Objects in Autonomous Vehicle 3D Sensing", 
    "pagination": "691-711", 
    "productId": [
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "pub.1141490051"
        ]
      }, 
      {
        "name": "doi", 
        "type": "PropertyValue", 
        "value": [
          "10.1007/978-3-030-88418-5_33"
        ]
      }
    ], 
    "publisher": {
      "name": "Springer Nature", 
      "type": "Organisation"
    }, 
    "sameAs": [
      "https://doi.org/10.1007/978-3-030-88418-5_33", 
      "https://app.dimensions.ai/details/publication/pub.1141490051"
    ], 
    "sdDataset": "chapters", 
    "sdDatePublished": "2022-09-02T16:10", 
    "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
    "sdPublisher": {
      "name": "Springer Nature - SN SciGraph project", 
      "type": "Organization"
    }, 
    "sdSource": "s3://com-springernature-scigraph/baseset/20220902/entities/gbq_results/chapter/chapter_110.jsonl", 
    "type": "Chapter", 
    "url": "https://doi.org/10.1007/978-3-030-88418-5_33"
  }
]
 

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.10.1007/978-3-030-88418-5_33'

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

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-88418-5_33'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-88418-5_33'

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

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-88418-5_33'


 

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

146 TRIPLES      22 PREDICATES      80 URIs      73 LITERALS      7 BLANK NODES

Subject Predicate Object
1 sg:pub.10.1007/978-3-030-88418-5_33 schema:about anzsrc-for:08
2 anzsrc-for:0801
3 schema:author N3c7d188dbabf437a8412873be29d752d
4 schema:datePublished 2021-09-30
5 schema:datePublishedReg 2021-09-30
6 schema:description LiDAR-driven 3D sensing allows new generations of vehicles to achieve advanced levels of situation awareness. However, recent works have demonstrated that physical adversaries can spoof LiDAR return signals and deceive 3D object detectors to erroneously detect “ghost" objects. Existing defenses are either impractical or focus only on vehicles. Unfortunately, it is easier to spoof smaller objects such as pedestrians and cyclists, but harder to defend against and can have worse safety implications. To address this gap, we introduce Shadow-Catcher, a set of new techniques embodied in an end-to-end prototype to detect both large and small ghost object attacks on 3D detectors. We characterize a new semantically meaningful physical invariant (3D shadows) which Shadow-Catcher leverages for validating objects. Our evaluation on the KITTI dataset shows that Shadow-Catcher consistently achieves more than 94% accuracy in identifying anomalous shadows for vehicles, pedestrians, and cyclists, while it remains robust to a novel class of strong “invalidation” attacks targeting the defense system. Shadow-Catcher can achieve real-time detection, requiring only between 0.003 s–0.021 s on average to process an object in a 3D point cloud on commodity hardware and achieves a 2.17x speedup compared to prior work.
7 schema:editor N46d6d253a5e34fc5975edfdbfbe17f68
8 schema:genre chapter
9 schema:isAccessibleForFree true
10 schema:isPartOf N665d5c689df1451cb5db2ad3cf74252b
11 schema:keywords KITTI dataset
12 accuracy
13 advanced level
14 adversary
15 attacks
16 awareness
17 class
18 cloud
19 commodity hardware
20 cyclists
21 dataset
22 defense
23 defense system
24 detection
25 detector
26 end
27 end prototype
28 evaluation
29 gap
30 generation
31 ghost objects
32 ghosts
33 hardware
34 implications
35 invalidation
36 invariants
37 levels
38 leverage
39 lidar
40 lidar return signals
41 new generation
42 new technique
43 novel class
44 object detector
45 objects
46 pedestrians
47 physical adversaries
48 physical invariants
49 point clouds
50 prior work
51 prototype
52 real-time detection
53 recent work
54 return signal
55 safety implications
56 sensing
57 set
58 shadow
59 signals
60 situation awareness
61 small objects
62 speedup
63 system
64 technique
65 vehicles
66 work
67 schema:name Shadow-Catcher: Looking into Shadows to Detect Ghost Objects in Autonomous Vehicle 3D Sensing
68 schema:pagination 691-711
69 schema:productId N434ba16226f94055ba62988793c9da52
70 Nc5e5019af5b14e889b80e4dd5aec5ab0
71 schema:publisher Nd53cd10c1f9f42fbb5e6dec2369a530a
72 schema:sameAs https://app.dimensions.ai/details/publication/pub.1141490051
73 https://doi.org/10.1007/978-3-030-88418-5_33
74 schema:sdDatePublished 2022-09-02T16:10
75 schema:sdLicense https://scigraph.springernature.com/explorer/license/
76 schema:sdPublisher N299a4d86ca9e476493ed910b474460f8
77 schema:url https://doi.org/10.1007/978-3-030-88418-5_33
78 sgo:license sg:explorer/license/
79 sgo:sdDataset chapters
80 rdf:type schema:Chapter
81 N0ab3b6d3fdae40a987a9bdae4077f7da rdf:first N71650f9d95b04ab3826d08b9f61af0e1
82 rdf:rest rdf:nil
83 N0ad8b6d5e6f14105ac5dcbdfebf5b7df rdf:first sg:person.013404167044.28
84 rdf:rest rdf:nil
85 N119edbe0e379409cadf460978b5db41e rdf:first sg:person.011365416202.34
86 rdf:rest Nc5f100b7987946eb9e97b0f8ae301840
87 N299a4d86ca9e476493ed910b474460f8 schema:name Springer Nature - SN SciGraph project
88 rdf:type schema:Organization
89 N3851179ef8404215af7ec2ad102945ad schema:familyName Bertino
90 schema:givenName Elisa
91 rdf:type schema:Person
92 N3a69c20db5174961bebab6590876b1e4 rdf:first N8924cb7672d94518bcb068610b9c1eac
93 rdf:rest N0ab3b6d3fdae40a987a9bdae4077f7da
94 N3c7d188dbabf437a8412873be29d752d rdf:first sg:person.014731402615.26
95 rdf:rest N119edbe0e379409cadf460978b5db41e
96 N434ba16226f94055ba62988793c9da52 schema:name doi
97 schema:value 10.1007/978-3-030-88418-5_33
98 rdf:type schema:PropertyValue
99 N46d6d253a5e34fc5975edfdbfbe17f68 rdf:first N3851179ef8404215af7ec2ad102945ad
100 rdf:rest N3a69c20db5174961bebab6590876b1e4
101 N665d5c689df1451cb5db2ad3cf74252b schema:isbn 978-3-030-88417-8
102 978-3-030-88418-5
103 schema:name Computer Security – ESORICS 2021
104 rdf:type schema:Book
105 N71650f9d95b04ab3826d08b9f61af0e1 schema:familyName Waidner
106 schema:givenName Michael
107 rdf:type schema:Person
108 N8924cb7672d94518bcb068610b9c1eac schema:familyName Shulman
109 schema:givenName Haya
110 rdf:type schema:Person
111 Nc5e5019af5b14e889b80e4dd5aec5ab0 schema:name dimensions_id
112 schema:value pub.1141490051
113 rdf:type schema:PropertyValue
114 Nc5f100b7987946eb9e97b0f8ae301840 rdf:first sg:person.011755172717.81
115 rdf:rest N0ad8b6d5e6f14105ac5dcbdfebf5b7df
116 Nd53cd10c1f9f42fbb5e6dec2369a530a schema:name Springer Nature
117 rdf:type schema:Organisation
118 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
119 schema:name Information and Computing Sciences
120 rdf:type schema:DefinedTerm
121 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
122 schema:name Artificial Intelligence and Image Processing
123 rdf:type schema:DefinedTerm
124 sg:person.011365416202.34 schema:affiliation grid-institutes:grid.7445.2
125 schema:familyName Demetriou
126 schema:givenName Soteris
127 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011365416202.34
128 rdf:type schema:Person
129 sg:person.011755172717.81 schema:affiliation grid-institutes:grid.7445.2
130 schema:familyName Muñoz-González
131 schema:givenName Luis
132 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011755172717.81
133 rdf:type schema:Person
134 sg:person.013404167044.28 schema:affiliation grid-institutes:grid.7445.2
135 schema:familyName Lupu
136 schema:givenName Emil C.
137 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013404167044.28
138 rdf:type schema:Person
139 sg:person.014731402615.26 schema:affiliation grid-institutes:grid.7445.2
140 schema:familyName Hau
141 schema:givenName Zhongyuan
142 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014731402615.26
143 rdf:type schema:Person
144 grid-institutes:grid.7445.2 schema:alternateName Imperial College London, London, UK
145 schema:name Imperial College London, London, UK
146 rdf:type schema:Organization
 




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


...