Sketch-based 3D model retrieval using compressive sensing classification

Yoon, Sungsoo; Kuijper, Arjan

doi:10.1049/el.2011.2158

Cited by 15 publications

(11 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first-tier measure is a well-known measure defined as the percentage of correctly retrieved objects within the top k retrieved objects in the respective query class, where k is the total number of relevant objects (i.e., the size of the query class). Figure 7 shows a comparison of the firsttier precision of our approach against the performance of the HELO in [36] and STELA descriptor as reported in [37], as well as the suggestive contour features without sparse coding representation and compressive sensing based classification as reported in [42]. Sparse coding-based feature optimization provides for improved first-tier precision in 8 of the 13 total classes, with significant improvements in some classes yielding up to 20 percentage points.…”

Section: Comparison Of Retrieval Performancementioning

confidence: 95%

“…We conducted experiments to evaluate the effectiveness of our proposed method as compared to two alternative retrieval methods [36,37] and one alternative coding approach [42]. We discuss the obtained results in three parts: the experimental setup in Section 4.1; the retrieval results from representative sketched query images by numerous users in Section 4.2; and comparison of our method against state-of-the-art previous approaches in Section 4.3.…”

Section: Methodsmentioning

confidence: 99%

“…In particular, the first tier precision of "chair" is more than double comparing to previous approaches because our proposed feature optimization Fig. 7 Comparison of user-drawn sketch based 3D model retrieval between our proposed approach and previous approaches HELO [36], STELA [37], and compressive sensing based classification [42] using sparse coding emphasize their own characteristics. However, the first tier precision of table is lower than previous approaches due to similar features of ant and octopus classes.…”

Section: Comparison Of Retrieval Performancementioning

confidence: 99%

See 2 more Smart Citations

User-drawn sketch-based 3D object retrievalusing sparse coding

Yoon

Schreck

2014

Multimed Tools Appl

View full text Add to dashboard Cite

3D object retrieval from user-drawn (sketch) queries is one of the important research issues in the areas of pattern recognition and computer graphics for simulation, visualization, and Computer Aided Design. The performance of any content-based 3D object retrieval system crucially depends on the availability of effective descriptors and similarity measures for this kind of data. We present a sketch-based approach for improving 3D object retrieval effectiveness by optimizing the representation of one particular type of features (oriented gradients) using a sparse coding approach. We perform experiments, the results of which show that the retrieval quality improves over alternative features and codings. Based our findings, the coding can be proposed for sketch-based 3D object retrieval systems relying on oriented gradient features. Keywords 3D object retrieval • Sketch-based querying • Gradient descriptor • Sparse coding 1 Introduction 3D object data is becoming ubiquitous in many application areas. It serves as the basis for computer aided design, where machining parts, buildings, or other types of objects are being composed of 3D object parts, to form a blueprint for real-world construction. 3D objects are also indispensable elements for scientific simulation, visualization, serious and

show abstract

Section: Comparison Of Retrieval Performancementioning

confidence: 95%

Section: Methodsmentioning

confidence: 99%

Section: Comparison Of Retrieval Performancementioning

confidence: 99%

See 1 more Smart Citation

User-drawn sketch-based 3D object retrievalusing sparse coding

Yoon

Schreck

2014

Multimed Tools Appl

View full text Add to dashboard Cite

show abstract

“…We then extract the suggestive contour [1], which is a robust estimate of the shape of the 3D model in a projected viewpoint, even if the surfaces of the 3D object are smooth. The suggestive contours are represented with a histogram of oriented gradients by analysing their properties in the topological space of diffusion tensor fields (HOG-DTFs) [2,3]. We need to optimise the HOG-DTF to remove noise and maintain the edges in order to robustly retrieve the 3D models.…”

mentioning

confidence: 99%

“…Fig. 3 shows the comparison of first-tier precision between the DSR-based approach [7], the HOG-DTF feature-based 3D model retrieval approach presented in [2], and the our method based on sparse coding. Sparse coding-based feature optimisation provides better first-tier precision in eight of the 13 total classes, with the improvement in some classes yielding up to 20 percentage points.…”

mentioning

confidence: 99%

Sparse coding based feature optimisation for robust 3D object retrieval

2012

Self Cite

View full text Add to dashboard Cite

3D object retrieval is an important research issue in the area of computer graphics for applications such as computer-aided design, simulation, and visualisation. The performance of 3D object retrieval systems is dependent on the development of efficient and effective descriptors and similarity measures. Presented is an effective approach for supporting 3D object retrieval based on optimising the gradient descriptor which has recently been shown to perform among the state of the art. The approach relies on sparse coding and the effectiveness of the method is demonstrated by experimental evaluation.Introduction: 3D object data is becoming ubiquitous in many application areas. It is basic to computer-aided design, in which machining parts, buildings, and other types of objects are composed of 3D object parts, forming blueprints for real-world construction. 3D objects are also indispensable elements in scientific simulation, visualisation, serious and entertainment gaming. To reuse existing 3D object content from large 3D object repositories, methods are needed for efficient retrieval of 3D objects as requested by users. Approaches for 3D object retrieval consider methods for implementing similarity functions that allow retrieval and clustering in large 3D object databases. To date, many retrieval methods have been proposed to compute the similarity between 3D objects for usage in retrieval algorithms. One of the most popular approaches is to extract descriptors for each 3D object and determine the similarity between each pair of objects based on the distance between them, calculated with a metric defined on the descriptors.In this Letter, we present an improved approach for supporting viewbased 3D object retrieval by applying feature optimisation in a higher dimension. We apply the approach on an encompassing benchmark data set representing various 3D objects that include pose, partial occlusion, and shape. We experimentally show that the proposed method outperforms an existing state-of-the-art method in terms of retrieval quality, and thus we recommend it for 3D object retrieval.

show abstract

Efficient Sketch-Based 3D Shape Retrieval via View Selection

2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Sketch-based 3D model retrieval using compressive sensing classification

Cited by 15 publications

References 7 publications

User-drawn sketch-based 3D object retrievalusing sparse coding

User-drawn sketch-based 3D object retrievalusing sparse coding

Sparse coding based feature optimisation for robust 3D object retrieval

Efficient Sketch-Based 3D Shape Retrieval via View Selection

Contact Info

Product

Resources

About