Search & Retrieval in CAD Databases

A User-centric State-of-the-Art Overview

Christoph Schinko

1,2

, Thomas Vosgien

, Thorsten Prante

, Tobias Schreck

and Torsten Ullrich

1,2

Fraunhofer Austria Research GmbH, Visual Computing @ Graz, Graz, Austria

Institute of Computer Graphics and Knowledge Visualization, Graz University of Technology, Graz, Austria

V-Research GmbH, Dornbirn, Austria

Keywords:

Computer-Aided Design, Information Retrieval, Library Services, CAD Search.

Abstract:

This article presents a state-of-the-art overview on shape, information and design retrieval systems in the

context of CAD engineering. In contrast to existing surveys, we classify the different approaches from a

CAD application user point of view. As a consequence, we focus on features of surveyed techniques such as:

supported shape data types, handling of geometric invariances, support of metadata, supported query types,

quality of retrieval results, and the availability of implementations.

1 MOTIVATION

The popularity of Computer-Aided Design (CAD)

and Product-Lifecycle-Management (PLM) systems

is based on a large amount of product information be-

ing generated and stored in engineering databases in

practice. This information is a source for new ways to

analyse and reuse existing designs. The reuse of de-

sign information is one of the important approaches to

increase design quality and productivity: An existing

design provides a basis for improvements incorporat-

ing the assemblies or components whose quality and

correctness have been proven in the past (Hoffmann

et al., 2014).

In engineering, it is conservatively estimated that

Figure 1: The top row shows a gimble ring, a lock nut, a

ﬂange and a washer (from left to right), which have a sim-

ilar shape but a different classiﬁcation according to their

function. The bottom row shows several washers (spring

lock washer, regular washer, hexagonal washer, knife-edge

washer; from left to right) with same name but different

shapes. – Image Source: (Jayanti et al., 2006) –

more than 75% of design activity comprises reuse of

previous design and manufacturing knowledge to ad-

dress a new design problem. Design reuse is achieved

by adaptation; i.e., the existing product is adapted to

a given requirement speciﬁcation. In order to make

a design reusable, the key requirement is an informa-

tion retrieval system. Although PDM systems have

greatly improved productivity in the design and man-

ufacture of new products, one of the disadvantages

with such systems is the possible difﬁculty of reusing

the product information stored by the system. Engi-

neers spend about 60% of their time searching for the

right information (Li et al., 2004).

In contrast to retrieval systems for multimedia

databases, 3D shape repositories or general pur-

poses, the engineering context reveals several chal-

lenges (Jayanti et al., 2006):

1. Engineering shapes are characterized by features

such as holes, tunnels, cavities, etc. The relative

position of these features are more important for a

part’s functionality than its overall shape.

2. The classiﬁcation of parts in the engineering con-

text has a low level of abstraction; e.g. a category

“airplanes” is not very reasonable in the context

of CAD, as an airplane would be considered as an

assembly of many much smaller objects.

3. In the CAD context, parts are often classiﬁed ac-

cording to their functionality and not according to

their geometric shape (cf. Figure 1).

306

Schinko C., Vosgien T., Prante T., Schreck T. and Ullrich T.

Search Retrieval in CAD Databases - A User-centr ic State-of-the-Art Overview.

DOI: 10.5220/0006268103060313

In Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2017), pages 306-313

ISBN: 978-989-758-224-0

2 RELATED WORK

2.1 Technical State of the Art

3D object retrieval is an important topic in product

lifecycle management. The various approaches to

tackle this problem have their roots in different re-

search domains: geometry processing, computer vi-

sion, pattern matching, knowledge management, etc.

Since almost two decades, many methods for 3D

shape description and retrieval have been researched.

Consequently, there is a need for surveys that system-

atize this research space. Within the last few years,

many state-of-the-art overviews have been published

and most of them are based on a taxonomy of the tech-

nique used. Table 1 lists important surveys in chrono-

logical order including bibliographic metrics.

Table 1: For each survey (in chronological order) this ta-

ble lists its number of pages, the number of included refer-

ences and the number of articles citing the survey. The ci-

tation numbers are based on the “Web of Science” database

by Thomson Reuters and on the “Microsoft Academia”

database (as of November 2016).

number

of pages

included

references

Web of

Science

Microsoft

Academia

(Loncaric, 1998) 19 223 – 1176

(Li et al., 2004) 9 79 – 6

(Zhang and Lu, 2004) 19 101 664 1866

(Bustos et al., 2005) 51 79 178 392

(Iyer et al., 2005) 22 154 201 556

(Bustos et al., 2006) 16 41 10 92

(Babic et al., 2008) 17 70 68 160

(Qin et al., 2008) 8 63 0 86

(Tangelder and Veltkamp, 2008) 31 136 265 1638

(Jayanti et al., 2009) 9 35 8 11

(Savelonas et al., 2015) 26 79 0 9

2.2 Benchmarks

Information retrieval in the context of Computer-

Aided Design and Product-Lifecycle-Management

should reﬂect a part’s functionality and not only its

geometric shape. General purpose shape benchmarks,

such as the Princeton Shape Benchmark (Shilane

et al., 2004), are only suited to a limited extent, as

they contain generic shapes.

The International Shape Retrieval Contest

(SHREC, http://www.shrec.net/) is a comprehensive

forum in which benchmarks and novel retrieval

challenges are deﬁned, with the goal to compare

retrieval algorithms. To date, the benchmark has ad-

dressed many different retrieval problems, including

global and local similarity, non-rigid similarity, or

multimodal retrieval. However, CAD retrieval to date

was not a core topic in SHREC.

A set of CAD-speciﬁc benchmarks have been pub-

lished by BESPALOV et al. in “Benchmarking CAD

Search Techniques”(Bespalov et al., 2005). A com-

prehensive engineering shape benchmark for CAD

models has been developed by JAYANTI et al. (Jayanti

et al., 2006). The benchmark is publicly available at:

engineering.purdue.edu

3 CAD MODEL RETRIEVAL

3.1 Information Retrieval

Information retrieval algorithms need to be evaluated

for the quality of provided retrieval results. To this

end, many measures are avialable (Baeza-Yates and

Ribeiro-Neto, 2008). Basic important measures to

rate a retrieval result include numbers of relevant and

retrieved documents (see Table 2).

A more expressive evaluation can be performed by us-

ing precision and recall. These measures are deﬁned

in terms of a set of retrieved documents and a set of

relevant documents. Precision p is the fraction of true

positive (tp) and the number of true positives plus the

number of false positives ( f p), whereas recall r is the

fraction of true positives and the number of true posi-

tives plus the number of false negatives ( fn):

p =

tp+ fp

, r =

tp+ fn

The two measurements are typically used to create

graphs – the precision/recall-plots. Since precision

can be seen as a measure of quality, and recall as a

measure of quantity, a curve at the top of the plot is

what systems are aiming at in information retrieval.

To evaluate measures likes precision and recall, ex-

ample queries and ground truth information needs to

be deﬁned in advance, e.g., by means of benchmarks

as mentioned in Section 2.2.

Table 2: Relevant and retrieved documents can give an in-

sight about the performance of an information retrieval sys-

tem.

relevant not relevant

retrieved true false

positives tp positives fp

not retrieved false true

negatives fn negatives tn

Search Retrieval in CAD Databases - A User-centric State-of-the-Art Overview

307

3.2 The Meaning of Shape

In the beginning of the 20. century MAX

WERTHEIMER began the formal founding of

Gestalt psychology (“Gestalt” is the German trans-

lation of “shape”, “form”) (King and Wertheimer,

2005), which attempts to understand the laws behind

the ability to acquire and maintain meaningful per-

ceptions. The central principle of gestalt psychology

maintains that when the human mind forms a percept,

the whole has a reality of its own, independent of

the parts.Conversely, IRVING BIEDERMAN explains

object recognition by a bottom-up process (Bieder-

man, 1987). According to his theory, we are able to

recognize objects by separating them into the object’s

main component parts–basic 3-dimensional shapes

such as cylinders, cones, etc. These basic shapes

form a simple “alphabet” that can be combined to

complex objects.

These principles have inﬂuenced many computer

graphics and computer vision techniques (Attene

et al., 2006). In “the meaning of shape and some

techniques to extract it” (Havemann et al., 2012) high-

light some of the fundamental but maybe less obvious

limitations of current methods for representing and

processing 3D shape. They introduce semantic en-

richment as central concept relating the subjective na-

ture of interpretation to the objective of classiﬁcation.

This semantic information, which describes an object

on a high, abstract level, is needed in order to provide

digital library services such as indexing, markup and

retrieval (Ullrich and Fellner, 2011). A digital library

provides services based on metadata. In the simplest

case, metadata is of the Dublin Core type (Dublin

Core Metadata Initiative, 1995) (title, creator/author,

time of creation, etc). This is insufﬁcient for large

databases with a huge number of 3D objects, because

of their versatility and rich structure.

As a consequence, automatic techniques for se-

mantic enrichment of CAD data is a vital research

topic. However, for many methods to calculate shape

signatures their relation to human perceptions of geo-

metric similarity is unknown (Clark et al., 2006).

3.3 Apples and Oranges

The problem of ill-deﬁned properties such as “geo-

metric similarity” can be illustrated even with a sim-

ple geometric shape: a circular sector. A feature vec-

tor to (completely) describe a circular sector may con-

sist of the disk radius r and the central angle α; for

example:

= (5.0cm, 30

◦

), c

= (5.5cm, 30

◦

), c

= (5.0cm, 32

◦

)

In order to measure “similarity” a metric is used

in many cases to deﬁne the distance between a

pair of feature vectors returning a non-negative

real number. A metric for two circular sectors

= (r

, α

) and c

= (r

, α

) is, for example,

d(c

, c

) =

− r

)

+ (α

− α

)

. Each metric is

faced with the problem to compare elements with

each other which have no natural order. In our exam-

ple, switching the speciﬁcation of the angles α from

degrees [0

◦

, 360

◦

] to radians [0, 2π] changes the order

of elements signiﬁcantly:

degrees : d(c

, c

) = 0.500 < 2.000 = d(c

, c

)

radians : d(c

, c

) = 0.500 > 0.035 ≈ d(c

, c

)

The effects of different distances using different met-

rics is illustrated in Figure 2. An overview on distance

and similarity measures can be found in the survey by

SUNG HYUK CHA (Cha, 2007).

α r

α α

Figure 2: The metric used to measure the distance between

feature vectors deﬁnes the “similarity” of shapes. Different

metrics (illustrated by different coordinate systems; left ↔

right) result in different distances and different similarities.

4 CLASSIFICATION

This state-of-the-art overview on shape and infor-

mation retrieval systems classiﬁes the different ap-

proaches to design retrieval from a CAD application

user’s point of view

. The classiﬁcation consists of

four categories inspecting different aspects of a digi-

tal library application.

4.1 Data and Metadata Representation

The ﬁrst category analyses the design retrieval ap-

proaches according to supported input data. Assum-

ing a user with an already existing CAD database, the

supported features within input data are an important

selection criterion. This criterion does not address

The classiﬁcation results are listed in the Tables 3

and 4. Despite every precaution these tables may contain

errors. If any errors are found, we ask your forgiveness, and

request you send us a short note pointing them out.

GRAPP 2017 - International Conference on Computer Graphics Theory and Applications

308

ﬁle formats but the geometric entities (point clouds,

polygonal meshes, NURBS, etc.) and the non geo-

metric entities (annotations, material properties, etc.)

which can be handled natively without conversion.

Geometric Input. Which geometric shape types and

model representations (point clouds, polygonal

data, analytical surfaces (NURBS, etc.), volumet-

ric data, constructive solid geometry (CSG), gen-

erative data) are supported natively without con-

version?

Manifoldness & Noise. Is the approach able to han-

dle inconsistencies, namely non-manifold geo-

metric input, and/or noisy input?

Level-of-Detail. If the database contains CAD mod-

els in different resolutions, does the inspected ap-

proach identify these “similar” models?

Invariants. Each approach uses a metric to identify

“similar” elements. Are a model and the same

model with an applied non-rigid/afﬁne/isometric

transformation similar according to the used met-

ric? In other words, is the used metric invariant to-

wards non-rigid/afﬁne/isometric transformations?

Metadata. Input data may contain metadata (mate-

rial properties, semantic annotations, copyright

information, etc.). Which kinds of metadata are

supported to be queried?

4.2 Queries and Results

Query Method. Which types of queries are sup-

ported (by example, by an image, by a taxonomy,

by a sketch, etc.)?

Subsets. Is it possible to query for a subpart only? In

other words, does the retrieval algorithm support

subpart matching?

Goodness. Are the queries and the corresponding re-

sults evaluated and benchmarked? And are the

benchmark results available?

4.3 Technology Readiness Level

Availability. Even though a retrieval algorithm may

be a signiﬁcant scientiﬁc contribution, without

an available implementation its beneﬁt from a

CAD application user’s point of view is limited.

What is the availability level of the presented ap-

proach (only article published, reference imple-

mentation available, library commercially/open-

source available, software released)?

Integration. If the implementation is available (com-

mercially or non-commercially), how can it be in-

tegrated into an existing CAD environment?

4.4 Technology

Method. A primary search method uses the (geo-

metric and non-geometric) content. A secondary

search method relies on a primary search method

and uses additional sources not contained in the in

the (geometric and non-geometric) content – such

as relevance-based user input. Is the analyzed ap-

proach a primary search method or a secondary

search method?

Performance. Are the results of computational

benchmarks considering speed and memory re-

quirements available?

5 CONCLUSIONS

This article presents a state-of-the-art overview on

shape, information and design retrieval systems in the

context of CAD engineering. In contrast to existing

surveys, we classify the different approaches from a

CAD application user point of view.

5.1 Open Problems

The classiﬁcation scheme in Section 4 has been de-

signed before the classiﬁcation took place. The

scheme is strongly inﬂuenced by interviews with in-

dustrial partners. Having classiﬁed the most im-

portant methods, some options of the classiﬁcation

scheme remain unused. Especially the availability of

implementations is limited. As a consequence, the

comparison of different approaches with each other is

complicated.

5.2 Beneﬁt

The presentation of an overview on state-of-the-art

techniques for CAD retrieval methods is beneﬁcial for

CAD users. In contrast to existing surveys, we focus

on critiria which are important from a CAD applica-

tion user point of view: supported shape data types,

handling of geometric invariances, support of meta-

data, supported query types, quality of retrieval re-

sults, and the availability of implementations.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the support

of the Austrian Research Promotion Agency, the

Forschungsf

orderungsgesellschaft (FFG) for the re-

search project AEDA (K-Projekt “Advanced Engi-

neering Design Automation”).

Search Retrieval in CAD Databases - A User-centric State-of-the-Art Overview

309

Table 3: This tabular overview lists the analyzed CAD retrieval approaches from a CAD application user’s point of view. All

approaches are ordered alphabetically according to the lastname of the ﬁrst author. This table is continued on the next page

(cf. Table 4).

Geometric Input

Non-Manifoldness

& Noise

Level of Detail

Invariants

Metadata

Query Method

Subpart Matching

Goodness

Availability

Integration

Method

Performance

(Abouel Nasr and Kamrani, 2006) Á Á À 3 À

(Akg

ul et al., 2009) Á 3 3 Á À 3 À 3

(Ansary et al., 2007b) Á 3 3 3 Á À 3 C++ À 3

(Ansary et al., 2007a) Á 3 3 3 Á À Á Â 3 C++ À 3

(Bai et al., 2010) Á Â À 3 À 3 3 À

(Bespalov et al., 2006) Á 3 3 Á À 3 3 À

(Biasotti et al., 2006) Á 3 Â À 3 3 À 3

(Cao et al., 2006) Á 3 Á Á 3 À

(Cardone et al., 2006) Á 3 Â À 3 3 À 3

(Chen et al., 2012) Á Â À 3 3 À 3

(Daras et al., 2006) Á 3 3 3 Á À 3 À 3

(Daras and Axenopoulos, 2009) Á 3 3 3 Á À Á Â 3 À

(Daras and Axenopoulos, 2010) Á 3 3 3 Á À Á Â 3 À 3

(Eitz et al., 2012) Á 3 3 3 Â Á 3 3 3 Python, C++ À

(El-Mehalawi and Miller, 2003b) Á Â Ã Á À 3 À

(El-Mehalawi and Miller, 2003a) Á Â Ã Á À 3 À

(Fisher and Hanrahan, 2010) Á 3 3 Â 3 Å 3 Á

(Fu et al., 2008) À Á Â À À 3 À

(Funkhouser et al., 2003) 3 Â 3 Á Ã Ä 3 3 C À 3

(Gao et al., 2006) Á Â 3 Â À 3 3 À

(Gao et al., 2010) Á Â 3 À À 3 À

(Getto and Fellner, 2015) Á 3 3 Â Ä 3 Á

(Grabner et al., 2014) À Á Â 3 3 3 À Ã Ä Á

(Grabner et al., 2015) À Á Â 3 3 3 À Ã Ä 3 Á 3

(Hong et al., 2006) Á Â 3 3 3 Á À 3 À

(Hou et al., 2005) Á Á À 3 Á

(Huangfu et al., 2016) Á Â À À 3 3 À 3

(Izadinia et al., 2016) Á Á Â À 3

(Kriegel et al., 2003) Á 3 3 Á À 3 À 3

(Kuo and Cheng, 2007) Á 3 3 Á À 3 À

(Leifman et al., 2005) Á 3 3 3 Á À 3 3 Á Á 3

(Leng and Qin, 2008) Á Á 3 À Ã 3 Á Á 3

(Leng and Xiong, 2009) Á Á 3 À Ã 3 Á Á 3

(Liu et al., 2006) Á À À 3 3 À 3

(Mademlis et al., 2006) Á Â Ã Ä 3 3 Á À 3 À 3

(Mademlis et al., 2009) Á Â Ã Ä 3 3 Á À 3 À 3

(Min et al., 2003) Á Á 3 À Á Ã À 3

GRAPP 2017 - International Conference on Computer Graphics Theory and Applications

310

Table 4: (continued from Table 3) This tabular overview lists the analyzed CAD retrieval approaches from a CAD application

user’s point of view. All approaches are ordered alphabetically according to the lastname of the ﬁrst author.

Geometric Input

Non-Manifoldness

& Noise

Level of Detail

Invariants

Metadata

Query Method

Subpart Matching

Goodness

Availability

Integration

Method

Performance

(Ohbuchi et al., 2005) Á 3 3 À À 3 À

(Papadakis et al., 2007) Á 3 Á À 3 À 3

(Papadakis et al., 2010) Á 3 Á À 3 À

(Pu et al., 2007) Á 3 3 À Á À 3

(Shih et al., 2007) Á 3 3 Â À 3 À 3

(Stavropoulos et al., 2010) Á 3 3 Á À 3 3 C++ À 3

(Sunil and Pande, 2008) Á 3 3 À À 3 3 À 3

(Tao et al., 2015) Á Â À À 3 3 À 3

(Vranic, 2005) Á 3 Á À 3 À 3

(Wang et al., 2008) Á Â 3 Á À Ã Ä 3 Á Á

(Yoon et al., 2010) Á 3 3 3 Á Á 3 À

(Zarpalas et al., 2006) Á Á À 3 À

(Zehtaban et al., 2016) Á Â À À Ä 3 À

Legend:

Geometrical Input À point clouds, Á polygonal data, Â analytical data, Ã CSG data, Ä volume data

Manifoldness & Noise 3, if non-manifold / noisy input data is supported

Level Of Detail 3, if level of detail matching is supported

Invariants metric is invariant towards À isometric, Á afﬁne, Â general non-rigid transformations

Metadata 3, if a content’s metadata is used by the matching algorithm

Query Method À by example, Á by sketch, Â by image, Ã by free text, Ä by taxonomy, Å by context

Subpart Matching 3, if it is possible to query for a subpart only

Goodness 3, if the presented approach is evaluated towards precision and recall

Availability 3, if the source code of the reference implementation is available

Integration the programming languages natively supported by the implementation

Method À a primary, content-based retrieval method, Á a secondary method on top of primary method

Performance 3, if the presented approach is evaluated towards computational time and memory consumption

REFERENCES

Abouel Nasr, E. S. and Kamrani, A. K. (2006). A new

methodology for extracting manufacturing features

from CAD system. Computers & Industrial Engineer-

ing, 51:389–415.

Akg

ul, C. B., Sankur, B., Yemez, Y., and Schmitt, F. (2009).

3d model retrieval using probability density-based

shape descriptors. IEEE Transactions on Pattern

Analysis and Machine Intelligence, 31:1117–1133.

Ansary, T. F., Daoudi, M., and Vandeborre, J.-P. (2007a). A

bayesian 3d search engine using adaptive views clus-

tering. IEEE Transactions on Multimedia, 9:78–88.

Ansary, T. F., Vandeborre, J.-P., and Daoudi, M. (2007b).

3d-model search engine from photos. Proceedings

of the ACM International Conference on Image and

Video Retrieval, 6:89–92.

Attene, M., Biasotti, S., Mortara, M., Patane, G., Spagn-

uolo, M., and Falcidieno, B. (2006). Computational

methods for understanding 3d shapes. Computers &

Graphics, 30:323–333.

Babic, B., Nesic, N., and Miljkovic, Z. (2008). A review of

automated feature recognition with rule-based pattern

recognition. Computers in Industry, 59:321–337.

Baeza-Yates, R. and Ribeiro-Neto, B. (2008). Modern

Information Retrieval. Addison-Wesley Publishing

Company, 2nd edition.

Bai, J., Goa, S., Tang, W., Liu, Y., and Guo, S. (2010). De-

sign reuse oriented partial retrieval of CAD models.

Computer-Aided Design, 42:1069–1084.

Bespalov, D., Ip, C. Y., Regli, W. C., and Shaffer, J. (2005).

Benchmarking CAD search techniques. Proceedings

of the ACM Symposium on Solid and Physical Model-

ing, 15:275–286.

Bespalov, D., Regli, W. C., and Shokoufandeh, A. (2006).

Search Retrieval in CAD Databases - A User-centric State-of-the-Art Overview

311

Local feature extraction and matching partial objects.

Computer-Aided Design, 38:1020–1037.

Biasotti, S., Marini, S., Spagnuolo, M., and Falcidieno,

B. (2006). Sub-part correspondence by structural

descriptors of 3d shapes. Computer-Aided Design,

38:1002–1019.

Biederman, I. (1987). Recognition-by-components: A the-

ory of human image understanding. Psychological Re-

view, 94(2):115–147.

Bustos, B., Keim, D., Saupe, D., Schreck, T., and Vrani

c, D.

(2006). An experimental effectiveness comparison of

methods for 3d similarity search. International Jour-

nal on Digital Libraries, 6:39–54.

Bustos, B., Keim, D. A., Saupe, D., Schreck, T., and Vrani

D. V. (2005). Feature-based similarity search in 3d

object databases. ACM Computing Surveys, 37:345–

387.

Cao, L., Liu, J., and Tang, X. (2006). 3d object retrieval

using 2d line drawing and graph based relevance feed-

back. Proceedings of the ACM International Confer-

ence on Multimedia, 14:105–108.

Cardone, A., Gupta, S. K., Deshmukh, A., and Karnik,

M. (2006). Machining feature-based similarity as-

sessment algorithms for prismatic machined parts.

Computer-Aided Design, 38:954–972.

Cha, S. H. (2007). Comprehensive survey on distance /

similarity measures between probability density func-

tions. Internaltional Journal of Mathematical Models

and Methods in Applied Sciences, 1:300–307.

Chen, X., Gao, S., and Guo, Song Bai, J. (2012). A

ﬂexible assembly retrieval approach for model reuse.

Computer-Aided Design, 44:554–574.

Clark, D. E. R., Corney, J. R., Mill, F., Rea, H. J., Sher-

lock, A., and Taylor, N. K. (2006). Benchmarking

shape signatures against human perceptions of geo-

metric similarity. Computer-Aided Design, 38:1038–

1051.

Daras, P. and Axenopoulos, A. (2009). A compact multi-

view descriptor for 3d object retrieval. Interna-

tional Workshop on Content-based Multimedia Index-

ing, 7:115–119.

Daras, P. and Axenopoulos, A. (2010). A 3d shape retrieval

framework supporting multimodal queries. Interna-

tional Journal of Computer Vision, 89:229–247.

Daras, P., Zarpalas, D., Tzovaras, D., and Strintzis, M. G.

(2006). Efﬁcient 3d model search and retrieval using

generalized 3d radon transforms. IEEE Transactions

on Multimedia, 8:101–114.

Dublin Core Metadata Initiative (1995). Dublin Core Meta-

data Initiative. http://dublincore.org/.

Eitz, M., Richter, R., Boubekeur, T., Hildebrand, K., and

Alexa, M. (2012). Sketch-based shape retrieval. ACM

Transactions on Graphics, 31:31–40.

El-Mehalawi, M. and Miller, R. A. (2003a). A database

system of mechanical components based on geometric

and topological similarity (part 2: Indexing, retrieval,

matching, and similarity assessment). Computer-

Aided Design, 35:95–105.

El-Mehalawi, M. and Miller, R. A. (2003b). A database

system of mechanical components based on geomet-

ric and topological similarity (part 1: Representation).

Computer-Aided Design, 35:83–94.

Fisher, M. and Hanrahan, P. (2010). Context-based search

for 3d models. ACM Transactions on Graphics,

29:182–191.

Fu, J., Joshi, S. B., and Simpson, T. W. (2008). Shape

differentiation of freeform surfaces using a similarity

measure based on an integral of gaussian curvature.

Computer-Aided Design, 40:311–323.

Funkhouser, T., Min, P., Kazhdan, M., Chen, J., Halderman,

A., Dobkin, D., and Jacobs, D. (2003). A search en-

gine for 3d models. ACM Transactions on Graphics,

22:83–105.

Gao, W., Gao, S. M., Liu, Y. S., Bai, J., and Hu, B. K.

(2006). Multiresolutional similarity assessment and

retrieval of solid models based on DBMS. Computer-

Aided Design, 38:985–1001.

Gao, Y., Dia, Q., and Zhang, N.-Y. (2010). 3d model com-

parison using spatial structure circular descriptor. Pat-

tern Recognition, 43:1142–1151.

Getto, R. and Fellner, D. W. (2015). 3d object retrieval with

parametric templates. Proceedings of the Eurograph-

ics Workshop on 3D Object Retrieval, 8:47–54.

Grabner, H., Ullrich, T., and Fellner, D. W. (2014). Content-

based retrieval of 3d models using generative mod-

eling techniques. Proceedings of EUROGRAPHICS

Workshop on Graphics and Cultural Heritage (Short

Papers / Posters), 12:9–12.

Grabner, H., Ullrich, T., and Fellner, D. W. (2015). Gen-

erative training for 3d retrieval. Proceedings of the

International Joint Conference on Computer Vision,

Imaging and Computer Graphics Theory and Appli-

cations (GRAPP), 10:97–105.

Havemann, S., Ullrich, T., and Fellner, D. W. (2012). The

meaning of shape and some techniques to extract it.

Multimedia Information Extraction, 1:81–98.

Hoffmann, C., Shapiro, V., and Srinivasan, V. (2014). Ge-

ometric interoperability via queries. Computer-Aided

Design, 46:148–159.

Hong, T., Lee, K., and Sungchan, K. (2006). Similarity

comparison of mechanical parts to reuse existing de-

signs. Computer-Aided Design, 38:973–984.

Hou, S., Lou, K., and Ramani, K. (2005). SVM-based se-

mantic clustering and retrieval of a 3d model database.

Computer-Aided Design and Applications, 2:155–

164.

Huangfu, Z.-M., Zhang, S.-S., and Yan, L.-H. (2016). A

method of 3d cad model retrieval based on spatial

bag of words. Multimedia Tools and Applications,

online:1–29.

Iyer, N., Jayanti, S., Lou, K., Kalyanaraman, Y., and Ra-

mani, K. (2005). Three-dimensional shape searching:

State-of-the-art review and future trends. Computer-

Aided Design, 37:509–530.

Izadinia, H., Shan, Q., and Seitz, S. M. (2016). IM2CAD.

arXiv preprint arXiv:1608.05137, 1:1–11.

Jayanti, S., Kalyanaraman, Y., Iyer, N., and Ramani, K.

(2006). Developing an engineering shape benchmark

for CAD models. Computer-Aided Design, 38:939–

953.

GRAPP 2017 - International Conference on Computer Graphics Theory and Applications

312

Jayanti, S., Kalyanaraman, Y., and Ramani, K. (2009).

Shape-based clustering for 3d CAD objects: A com-

parative study of effectiveness. Computer-Aided De-

sign, 41:999–1007.

King, B. D. and Wertheimer, M. (2005). Max Wertheimer

& Gestalt Theory. Transaction Publishers. ISBN 0-

7658-0258-9.

Kriegel, H.-P., Brecheisen, S., Kr

oger, P., Pfeiﬂe, M., and

Schubert, M. (2003). Using sets of feature vectors for

similarity search on voxelized CAD objects. Proceed-

ings of the ACM International Conference on Man-

agement of Data, 30:587–598.

Kuo, C.-T. and Cheng, S.-C. (2007). 3d model retrieval

using principal plane analysis and dynamic program-

ming. Pattern Recognition, 40:742–755.

Leifman, G., Meir, R., and Tal, A. (2005). Semantic-

oriented 3d shape retrieval using relevance feedback.

The Visual Computer, 21:865–875.

Leng, B. and Qin, Z. (2008). A powerful relevance feed-

back mechanism for content-based 3d model retrieval.

Multimedia Tools and Applications, 40:135–150.

Leng, B. and Xiong, Z. (2009). ModelSeek: An effective

3d model retrieval system. Multimedia Tools and Ap-

plications, 51:935–962.

Li, Z., Liu, M., and Ramani, K. (2004). Review of prod-

uct information retrieval: Representation and index-

ing. Proceedings of Design Engineering Technical

Conferences (DETC), 4:971–979.

Liu, Y., Zha, H., and Qin, H. (2006). Shape topics: A com-

pact representation and new algorithms for 3d partial

shape retrieval. IEEE Computer Society Conference

on Computer Vision and Pattern Recognition, 2:2025–

2032.

Loncaric, S. (1998). A survey of shape analysis techniques.

Pattern Recognition, 31:983–1001.

Mademlis, A., Axenopoulos, A., Daras, P., Tzovaras, D.,

and Strintzis, M. G. (2006). 3d content-based search

based on 3d krawtchouk moments. Proceedings of the

International Symposium on 3D Data Processing, Vi-

sualization, and Transmission, 3:743–749.

Mademlis, A., Daras, P., Tzovaras, D., and Strintzis, M. G.

(2009). 3d object retrieval using the 3d shape impact

descriptor. Pattern Recognition, 42:2447–2459.

Min, P., Halderman, J. A., Kazhdan, M., and Funkhouser,

T. A. (2003). Early experiences with a 3d model

search engine. Proceedings of the International Con-

ference on 3D Web Technology, 8:7–18.

Ohbuchi, R., Minamitani, T., and Takei, T. (2005). Shape-

similarity search of 3d models by using enhanced

shape functions. International Journal of Computer

Applications in Technology, 23:70–85.

Papadakis, P., Pratikakis, I., Perantonis, S., and Theoharis,

T. (2007). Efﬁcient 3d shape matching and retrieval

using a concrete radialized spherical projection repre-

sentation. Pattern Recognition, 40:2437–2452.

Papadakis, P., Pratikakis, I., Theoharis, T., and Peranto-

nis, S. (2010). PANORAMA: A 3d shape descriptor

based on panoramic views for unsupervised 3d object

retrieval. International Journal of Computer Vision,

89:177–192.

Pu, J., Kalyanaraman, Y., Jayanti, Subramaniam Ramani,

K., and Pizlo, Z. (2007). Navigation and discovery in

3d CAD repositories. IEEE Computer Graphics and

Applications, 27:38–47.

Qin, Z., Jia, J., and Qin, J. (2008). Content based 3d

model retrieval: A survey. International Workshop on

Content-based Multimedia Indexing, 6:249–256.

Savelonas, M. A., Pratikakis, I., and Sﬁkas, K. (2015). An

overview of partial 3d object retrieval methodologies.

Multimedia Tools and Applications, 74:11783–11808.

Shih, J.-L., Lee, C.-H., and Wang, J. T. (2007). A new 3d

model retrieval approach based on the elevation de-

scriptor. Pattern Recognition, 40:283–295.

Shilane, P., Min, P., Kazhdan, M., and Funkhouser, T. A.

(2004). The princeton shape benchmark. Shape Mod-

eling International, 8:1–12.

Stavropoulos, G., Moschonas, P., Moustakas, K., Tzovaras,

D., and Strintzis, M. G. (2010). 3d model search

and retrieval from range images using salient features.

IEEE Transactions on Multimedia, 12:692–704.

Sunil, V. B. and Pande, S. S. (2008). Automatic recogni-

tion of features from freeform surface CAD models.

Computer-Aided Design, 40:502–517.

Tangelder, J. W. H. and Veltkamp, R. C. (2008). A survey of

content based 3d shape retrieval methods. Multimedia

Tools and Applications, 39:441–471.

Tao, S., Wang, S., and Chen, A. (2015). 3d cad solid model

retrieval based on region segmentation. Multimedia

Tools and Applications, online:1–19.

Ullrich, T. and Fellner, D. W. (2011). Generative object

deﬁnition and semantic recognition. Proceedings of

the Eurographics Workshop on 3D Object Retrieval,

4:1–8.

Vranic, D. V. (2005). DESIRE: A composite 3d shape de-

scriptor. IEEE International Conference on Multime-

dia and Expo, 6:1–4.

Wang, X., Lv, T., Wang, S., and Wang, Z. (2008). An ontol-

ogy and SWRL based 3d model retrieval system. Asia

Information Retrieval Symposium (LNCS), 4993:335–

344.

Yoon, S. M., Scherer, M., Schreck, T., and Kuijper, A.

(2010). Sketch-based 3d model retrieval using diffu-

sion tensor ﬁelds of suggestive contours. Proceedings

of the ACM International Conference on Multimedia,

18:193–200.

Zarpalas, D., Daras, P., Axenopoulos, A., Tzovaras, D., and

Strintzis, M. G. (2006). 3d model search and retrieval

using the spherical trace transform. EURASIP Journal

on Advances in Signal Processing, 2007:1–14.

Zehtaban, L., Elazhary, O., and Roller, D. (2016). A

framework for similarity recognition of CAD models.

Journal of Computational Design and Engineering,

3:274–285.

Zhang, D. and Lu, G. (2004). Review of shape representa-

tion and description techniques. Pattern Recognition,

37:1–19.

Search Retrieval in CAD Databases - A User-centric State-of-the-Art Overview

313