Title |
SEGMENTATION AND ALIGNMENT OF MULTI-ORIENTED AND CURVED TEXT LINES FROM DOCUMENT IMAGES |
| Int J Mach Intell Vol:6 Iss:1 (2015-06-04) : 426-434 |
Authors |
T.K. BOAZ, C.J. PRABHAKAR |
Published on |
04 Jun 2015 Pages : 426-434 Article Id : BIA0002481 Views : 992 Downloads : 722 |
|
Abstract |
Full Text |
PDF | XML |
PubMed XML |
CNKI |
Cited By |
Open Access | Research Article
In this paper, we present a novel approach to segment and align multi-oriented and curved text-lines from document images. We assumed that the input document image contains text-lines with arbitrary orientation and identified the arbitrary text string based on projection profile. We employed anisotropic Gaussian filter bank on the identified arbitrary text region in order to smooth the text region, which helps to detect the ridges which is a representative of a text-line path. The ridges are then labeled and a cubic B-spline is fitted to the text-line path points. The orientation and curvature features of the text-line path is estimated using orientated gradients for each point and corresponding curvature to these text-line path are computed. Text is aligned along the horizontally transformed line by rotating individual characters based on the computed curvature information. Finally, the aligned text-lines are extracted, which can be fed into OCR for recognition. The evaluation metrics was evaluated at text-line segmentation level and the results posted show a significant improvement. The resulting system is proven to provide better results than most state of the art algorithms.
|
|
Title |
SCHEDULING ALGORITHM FOR QUALITY-OF-SERVICE DIFFERENTIATION IN A MIXTURE OF REAL-TIME AND NON-REAL-TIME TRAFFICS |
| Int J Mach Intell Vol:6 Iss:1 (2015-06-10) : 435-440 |
Authors |
O. BELLO, H. ZEN, A.K. OTHMAN, K. AB-HAMID |
Published on |
10 Jun 2015 Pages : 435-440 Article Id : BIA0002483 Views : 991 Downloads : 738 |
|
Abstract |
Full Text |
PDF | XML |
PubMed XML |
CNKI |
Cited By |
Open Access | Research Article
In this paper, we propose a scheduling algorithm to differentiate quality-of-service (QoS) and improve the users’ QoS satisfaction in heterogeneous mixed traffic scenarios. To achieve these, we develop a scheduling algorithm based on three novel radio resource allocation (RRA) techniques; delay-based scheduling policy for real-time (RT) services using a sigmoid-like utility function, minimum-rate-based scheduling policy for non-real-time (NRT) services which obeys the law of diminishing marginal utility and a throughput-based scheduling policy for best-effort (BE) which is also based on the law of diminishing marginal utility but without a minimum rate requirement. Simulation study shows superior performances of the proposed scheduling algorithm compared to some existing ones in terms of system throughput and user satisfaction both in single traffic and heterogeneous mixed traffic scenarios.
|
|
Title |
A NOVEL FEATURE LEVEL FUSION APPROACH USING FINGERPRINT BIOMETRICS FOR GENDER CLASSIFICATION |
| Int J Mach Intell Vol:6 Iss:1 (2015-06-15) : 441-444 |
Authors |
S.S. REVATE, P.D. DESHMUKH, K.V. KALE |
Published on |
15 Jun 2015 Pages : 441-444 Article Id : BIA0002495 Views : 998 Downloads : 747 |
|
Abstract |
Full Text |
PDF | XML |
PubMed XML |
CNKI |
Cited By |
Open Access | Research Article
Gender classification from fingerprints is an important step in forensic anthropology in order to identify the gender of a criminal and minimize the list of suspects search. Wavelets have been used frequently in image processing, feature extraction, compression and image super-resolution. Singular value decomposition (SVD) is quite possibly the most widely-used multivariate statistical technique used in the atmospheric sciences. Â A novel method of gender Classification from fingerprint is proposed based on Stationary Wavelet Transform (SWT) and singular value decomposition (SVD). The classification Stationary Wavelet Transform is achieved by extracting the energy computed from all the sub-bands of Stationary Wavelet Transform combined with singular values obtained from the SVD of fingerprint images. K nearest neighbor (KNN) used as a classifier.
|
|
Title |
A EXPERT SYSTEM FOR QUALITY ASSESSMENT AND ACCREDITATION (ESQAA) IN HIGHER EDUCATIONAL INSTITUTES |
| Int J Mach Intell Vol:6 Iss:1 (2015-06-18) : 445-449 |
Authors |
A.T. GAIKWAD, R.V. KULKARNI |
Published on |
18 Jun 2015 Pages : 445-449 Article Id : BIA0002496 Views : 976 Downloads : 735 |
|
Abstract |
Full Text |
PDF | XML |
PubMed XML |
CNKI |
Cited By |
Open Access | Research Article
An Expert System is a Computer program designed to solve problems in a domain in which there is human expertise. The Knowledge built into the system is usually obtained from experts in the field the field may be any from business applications or from technical background. By using this knowledge, an expert system can replicate the thinking process of the human experts. The experts are not common and they are rare. We have designed and developed a rule based expert system for quality assessment and enhancement in higher educational Institutes.
This paper deals with the development of an expert system for quality assessment and accreditation in Higher educational Institutes. Expert systems have been useful in performing experts task in various areas of medicine, engineering, research and business applications. The use of expert system for higher education is new and it has innovative contribution in the field of research. We have designed the expert system for quality assessment and enhancement in higher educational Institutes using NAAC Quality parameters. The expert system has User interface, knowledgebase to store the various parameters of quality suggested by NAAC. We have designed the rules to generate the grade of the Institutes. The interface which takes various quality parameters is designed using VB-6. The Expert System developed have given short acronym as ESQAA. It is useful to the higher educational Institutes to know the present quality status and various deficiencies as per the criteria suggested by NAAC and systematic suggestions to improve the grade. This is an decent contribution in improving the quality in higher educational institutions. We have given only sample outputs and certain rules to reduce the length of the research paper.
|
|
Title |
MINING FREQUENT ITEMSETS FROM DISTRIBUTED DATA USING DIC-LIKE APPROACH |
| Int J Mach Intell Vol:6 Iss:1 (2015-06-29) : 450-453 |
Authors |
R.N. YADAWAD, R.B.V. SUBRAMANYAM, U.P. KULKARNI, K. KORI |
Published on |
29 Jun 2015 Pages : 450-453 Article Id : BIA0002497 Views : 1003 Downloads : 818 |
|
Abstract |
Full Text |
PDF | XML |
PubMed XML |
CNKI |
Cited By |
Open Access | Research Article
DIC algorithm is an Apriori-like algorithm proposed for mining frequent itemsets from data. In this paper, an algorithm is designed to find frequent itemsets from data stored in a distributed system. Motivation to the proposed approach is from studying the applicability of DIC algorithm to distributed system. Issues sorted in the process of tailoring DIC include (1) Indicating right time for communication of partial results to other systems for their optimal extraction of frequent itemsets of higher dimension; (2) Optimal portion of data or results to be transmitted over the network keeping the goal of minimal bandwidth consumption. In literature, some algorithms are proposed for mining global frequent itemsets in a distributed system. Number of scans over the databases, amount of communication needed and overall computation time required, are some issues to improve further.
|