• Proceedings of the IEEK Conference
• /
• 2002.07c
• /
• pp.1420-1423
• /
• 2002

The classification and recognition of two-dimensional trademark patterns independently of their position, orientation, size and scale by proposing two feature vectors has been discussed. The paper presents experimentation on two feature vectors showing size- invariance and scale-invariance respectively. Both feature vectors are equally invariant to rotation as well. The feature extraction is based on local as well as global statistics of the image. These feature vectors have appealing mathematical simplicity and are versatile. The results so far have shown the best performance of the developed system based on these unique sets of feature. The goal has been achieved by segmenting the image using connected-component (nearest neighbours) algorithm. Second part of this work considers the possibility of using back propagation neural networks (BPN) for the learning and matching tasks, by simply feeding the feature vectosr. The effectiveness of the proposed feature vectors is tested with various trademarks, not used in learning phase.

• Journal of Information Processing Systems
• /
• v.8 no.1
• /
• pp.85-100
• /
• 2012

The Hartley transform based feature extraction method is proposed for fingerprint matching. Hartley transform is applied on a smaller region that has been cropped around the core point. The performance of this proposed method is evaluated based on the standard database of Bologna University and the database of the FVC2002. We used the city block distance to compute the similarity between the test fingerprint and database fingerprint image. The results obtained are compared with the discrete wavelet transform (DWT) based method. The experimental results show that, the proposed method reduces the false acceptance rate (FAR) from 21.48% to 16.74 % based on the database of Bologna University and from 31.29% to 28.69% based on the FVC2002 database.

• Journal of Information Processing Systems
• /
• v.13 no.4
• /
• pp.852-862
• /
• 2017

In this work a Discrete Cosine Transform (DCT)-based feature dimensionality reduced approach for fingerprint matching is proposed. The DCT is applied on a small region around the core point of fingerprint image. The performance of our proposed method is evaluated on a small database of Bologna University and two large databases of FVC2000. A dimensionally reduced feature vector is formed using only approximately 19%, 7%, and 6% DCT coefficients for the three databases from Bologna University and FVC2000, respectively. We compared the results of our proposed method with the discrete wavelet transform (DWT) method, the rotated wavelet filters (RWFs) method, and a combination of DWT+RWF and DWT+(HL+LH) subbands of RWF. The proposed method reduces the false acceptance rate from approximately 18% to 4% on DB1 (Database of Bologna University), approximately 29% to 16% on DB2 (FVC2000), and approximately 26% to 17% on DB3 (FVC2000) over the DWT based feature extraction method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.12
• /
• pp.3149-3165
• /
• 2013

Content based image retrieval is increasingly gaining popularity among image repository systems as images are a big source of digital communication and information sharing. Identification of image content is done through feature extraction which is the key operation for a successful content based image retrieval system. In this paper content based image retrieval system has been developed by adopting a strategy of combining multiple features of shape, color and relevance feedback. Shape is served as a primary operation to identify images whereas color and relevance feedback have been used as supporting features to make the system more efficient and accurate. Shape features are estimated through second derivative, least square polynomial and shapes coding methods. Color is estimated through max-min mean of neighborhood intensities. A new technique has been introduced for relevance feedback without bothering the user.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.30 no.2
• /
• pp.143-152
• /
• 2012

It is well-known that image matching is necessary for automatic generation of 3D data such as digital surface data from aerial images. Recently developed aerial digital cameras allow to capture multi-strip images with higher overlaps and less occluded areas than conventional analogue cameras and that much of researches on multi-image matching have been performed, particularly effective methods of measuring a similarity among multi-images using point features as well as linear features. This research aims to investigate similarity measuring methods such as SSD and SNCC incorporated into a area based multi-image matching method based on vertical line locus. In doing this, different similarity measuring entities such as grey value, grey value gradient, and average of grey value and its gradient are implemented and analyzed. Further, both dynamic and pre-fixed adaptive-window size are tested and analyzed in their behaviors in measuring similarity among multi-images. The aerial images used in the experiments were taken by a DMC aerial frame camera in three strips. The over-lap and side-lap are about 80% and 60%, respectively. In the experiment, it was found that the SNCC as similarity measuring method, the average of grey value and its gradient as similarity measuring entity, and dynamic adaptive-window size can be best fit to measuring area-based similarity in area based multi-image matching method based on vertical line locus.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.306-309
• /
• 1996

This paper presents a method of landmark recognition in indoor environments using a neural-network for an autonomous mobile robot. In order to adapt to image deformation of a landmark resulted from variations of view-points and distances, a multi-labeled template matching(MLTM) method and a dynamic area search method(DASM) are proposed. The MLTM is. used for matching an image template with deformed real images and the DASM is proposed to detect correct feature points among incorrect feature points. Finally a feed-forward neural-network using back-propagation algorithm is adopted for recognizing the landmark.

• Journal of the Semiconductor & Display Technology
• /
• v.8 no.4
• /
• pp.95-100
• /
• 2009

This paper presents an efficient disparity matching, using sum of absolute difference (SAD) and dynamic programming (DP) algorithm. This algorithm makes use of one of area-based algorithm which is the absolute sum of the pixel difference corresponding to the window size. We use the information of the right eye brightness (B) and the left eye brightness to get an best matching results and apply the results to the left eye image using the window go by the brightness of the right eye image. This is that we can control the brightness. The major feature of this algorithm called SAD+DP+B is that although Root Mean Square (RMS) performance is slightly less than SAD+DP, due to comparing original image, its visual performance is increased drastically for matching the disparity map on account of its matching compared to SAD+DP. The simulation results demonstrate that the visual performance can be increased and the RMS is competitive with or slightly higher than SAD+DP.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.3
• /
• pp.894-907
• /
• 2022

The features extracted by convolutional neural networks are more descriptive of images than traditional features, and their convolutional layers are more suitable for retrieving images than are fully connected layers. The convolutional layer features will consume considerable time and memory if used directly to match an image. Therefore, this paper proposes a feature weighting and region integration method for convolutional layer features to form global feature vectors and subsequently use them for image matching. First, the 3D feature of the last convolutional layer is extracted, and the convolutional feature is subsequently weighted again to highlight the edge information and position information of the image. Next, we integrate several regional eigenvectors that are processed by sliding windows into a global eigenvector. Finally, the initial ranking of the retrieval is obtained by measuring the similarity of the query image and the test image using the cosine distance, and the final mean Average Precision (mAP) is obtained by using the extended query method for rearrangement. We conduct experiments using the Oxford5k and Paris6k datasets and their extended datasets, Paris106k and Oxford105k. These experimental results indicate that the global feature extracted by the new method can better describe an image.

• Journal of Korea Multimedia Society
• /
• v.19 no.2
• /
• pp.138-145
• /
• 2016

LLAH(Locally Likely Arrangement Hashing) is a method which describes image features by exploiting the geometric relationship between their neighbors. Inherently, it is more robust to large view change and poor scene texture than conventional texture-based feature description methods. However, LLAH strongly requires that image features should be detected with high repeatability. The problem is that such requirement is difficult to satisfy in real applications. To alleviate the problem, this paper proposes a method that improves the matching rate of LLAH by exploiting together the brightness of features. Then, it is verified that the matching rate is increased by about 5% in experiments with synthetic images in the presence of Gaussian noise.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.2 s.40
• /
• pp.159-167
• /
• 2006

Lipreading has been suggested as one of the methods to improve the performance of speech recognition in noisy environment. However, existing methods are developed and implemented only in software. This paper suggests a hardware design for real-time lipreading. For real-time processing and feasible implementation, we decompose the lipreading system into three parts; image acquisition module, feature vector extraction module, and recognition module. Image acquisition module capture input image by using CMOS image sensor. The feature vector extraction module extracts feature vector from the input image by using parallel block matching algorithm. The parallel block matching algorithm is coded and simulated for FPGA circuit. Recognition module uses HMM based recognition algorithm. The recognition algorithm is coded and simulated by using DSP chip. The simulation results show that a real-time lipreading system can be implemented in hardware.