Applying an Artificial Neural Network to Pattern Recognition in Fabric Defects

Tsai, I-Shou; Lin, Chung-Hua; Lin, Jeng-Jong

doi:10.1177/004051759506500301

Cited by 100 publications

(59 citation statements)

References 4 publications

(11 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Many attempts have been made in order to perform the inspection automatically. Consequently the task of automated defects detection is popular and many research teams have focused their interest on it, while many of them have used ANNs to support the fault detection task, (Tsai et al, 1995;Sette & Bullard, 1996;Tilocca et al, 2002, Kumar, 2003Islam et al, 2006;Shady et al, 2006;Behera & Mani, 2007;Mursalin at al., 2008). Another similar approach is the combined use of fuzzy systems (Choi et al, 2001;Huang & Chen, 2003) or wavelet packet bases (Hu & Tsai, 2000;Jianli & Baoqi, 2007).…”

Section: Fabricsmentioning

confidence: 99%

Artificial Neural Networks and Their Applications in the Engineering of Fabrics

Vassiliadis¹,

Rangoussi²,

Çay³

et al. 2010

Woven Fabric Engineering

View full text Add to dashboard Cite

Section: Fabricsmentioning

confidence: 99%

Artificial Neural Networks and Their Applications in the Engineering of Fabrics

Vassiliadis¹,

Rangoussi²,

Çay³

et al. 2010

Woven Fabric Engineering

View full text Add to dashboard Cite

“…25 Behalwe vir reël-gefundeerde en geval-gefundeerde redenasiestelsels mag ekspertstelsels ook ander kennisverteenwoordigende metodes insluit, soos raamwerke, semantiese netvorming, neurale netwerke en "wasiglogika". In tekstiele, is neurale netwerke gebruik om materiaalfoute te identifi seer, 17,27 om drapering van kledingstukke te voorspel, 13 materiaal te ontwikkel 12 en wolgaringlasse 19 te klassifi seer.…”

Section: Verwante Werkunclassified

A distributed knowledge-based system for the optimum utilisation of South African wool

Dlodlo

Hunter

Botha³

et al. 2009

Suid-Afrikaanse Tydskr Natuurwetenskap Tegnol

View full text Add to dashboard Cite

This article describes the concept and development of a knowledge-based advisory system for the optimum utilisation of South African wool for the beneﬁt of present and potential investors and other interested parties. Wool is a natural animal ﬁbre produced in varying quantities around the world. The wool ﬁbre is far from homogenous; its type and quality, such as ﬁneness and length, depending on the breed of sheep and the environmental conditions prevailing during its growth. Wool is used in a variety of end uses, ranging from ﬁ ne worsted suiting, to hand knitting yarn, carpets, blankets and aircraft upholstery, its use depending largely on its ﬁbre ﬁneness and length. The wool industry is one of the oldest agricultural industries in South Africa, playing an important economic role as an earner of foreign exchange, and providing a living to many people. Wool is produced in many parts of South Africa under extensive, semi-extensive or intensive conditions, and is largely an export commodity. It is produced and traded in a sophisticated free market business environment into the international market place, where supply and demand forces determine price levels. More than 90% of locally produced wool is exported in an unprocessed or semi-processed form which detrimentally affects employment, foreign exchange and income-generating opportunities associated with value-addition prior to export. To reduce the amount of wool exported in unprocessed or semi-processed form, wool-processing enterprises need to be established to produce internationally marketable end products. Therefore, South Africa needs to attract investors into the wool sector, who will set up manufacturing mills in an economically sustainable manner. Potential and present investors in the South African (S.A.) wool industry need easily accessible and up-to-date information on the production statistics, processing properties and end-use pplications of the wool they need for the particular end-products they manufacture or could manufacture. To achieve this and ensure accessibility to such continuously updated information, it is essential to develop an integrated computer-based system. It is with the above in mind that a knowledge-based system for the optimum utilisation of South African wool has been developed, which is described here. This paper reviews relevant work in this ﬁ eld and covers wool production statistics in South Africa, the end uses of the wool ﬁbre versus the diameter of the ﬁbre, the advantages of distributed architectures, and the ﬂow of processes in a wool utilization system. It then sets out the concept and development of the proposed system, including the architecture of the proposed expert system, the associated analysis and ﬁnally the conclusions. The components of the expert system, namely the knowledge base, inference engine, knowledge acquisition component, and explanation system are described. The architecture of the system incorporates the concept of distributed systems and the related advantages incorporated in its general architecture and within its internal components. It marries both expert and general knowledge-based systems, consisting of a combination of an ordinary knowledge-based system (KBS) that can be queried for information and an expert system that provides advice to users. The distributed system developed involves collection of autonomous components that are interconnected, which enables these components to coordinate their activities and share resources of the system, so that users perceive the system as a single integrated facility. There are a number of advantages of such a distributed system and these are articulated in the paper. This approach allows not only incremental development of the system, but also facilitates sharing of data and information. The distributed nature of the architecture of the system developed, consists of three main elements: The expert system to advise on the characteristics of the wool that is required for a particular end use A knowledge-based system for querying on the distribution of wool of the various characteristics in South Africa An expert system for the selection of the best alternative area for investment for the particular product end use.The knowledge base consists of a number of databases, each representing the various wool characteristics. This represents a distributed architecture of the knowledge base. Therefore, this architecture inherits all the advantages of distributed processing systems as described in the paper. These knowledge bases can be queried by the user via a database management system (DBMS), a software that manages the creation, updating, maintenance and querying of the database. In terms of wool utilization, the system involves capturing the end-use and requirements of a product and from it, retrieving the characteristics of the wool that will meet the particular end-use. The availability of the wool is then checked by region and province for each style, type, clip type, yield, colour, vegetable matter fault and micron range, in line with the latest statistics available.The system developed enables questions such as the following to be asked at the user interface: What is the anticipated end use of the wool? What criteria must the wool satisfy for the selected end-use? What quantities of wool are required?The outputs at the user interface of the system are the quantities of wool per province and region in terms of micron, style, yield, colour, type, clip type as available on the web-site of Cape Wools SA. At the very end of the system, the best alternative site for siting the manufacturing base can also be indicated.

show abstract

“…Statistical approaches [27,28,36] , using diverse statistical properties of textures and defects, may effectively detect fabric defects. But the texture pattern and defect shape have a crucial influence on the detection results of statistical and spectral approaches, such as [27,30], see Fig.…”

Section: Introductionmentioning

confidence: 99%

Fabric defect inspection using prior knowledge guided least squares regression

Cao

Zhang

Wen

et al. 2015

Multimed Tools Appl

View full text Add to dashboard Cite

This paper proposes an unsupervised model to inspect various detects in fabric images with diverse textures. A fabric image with defects is usually composed of a relatively consistent background texture and some sparse defects, which can be represented as a low-rank matrix plus a sparse matrix in a certain feature space. The process is formulated as a least squares regression based subspace segmentation model, which is convex, smooth and can be solved efficiently. A simple and effective prior is also learnt from local texture features of the image itself. Instead of considering only the feature space' s global structure, the local prior is incorporated with it seamlessly by the proposed subspace segmentation model to guide and improve the segmentation. Experiments on a variety of fabric images demonstrate the effectiveness and robustness of the proposed method. Compared with existing methods, our method is more robust and locates various defects more precisely.

show abstract

Applying an Artificial Neural Network to Pattern Recognition in Fabric Defects

Cited by 100 publications

References 4 publications

Artificial Neural Networks and Their Applications in the Engineering of Fabrics

Artificial Neural Networks and Their Applications in the Engineering of Fabrics

A distributed knowledge-based system for the optimum utilisation of South African wool

Fabric defect inspection using prior knowledge guided least squares regression

Contact Info

Product

Resources

About