• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.10a
• /
• pp.196-199
• /
• 2010

The human eyes are important facial landmarks for image normalization due to their relatively constant interocular distance. This paper introduces a novel approach for the eye detection task using optimal segmentation method for eye representation. The method consists of three steps: (1)edge extraction method that can be used to accurately extract eye region from the gray-scale face image, (2)extraction of eye region using labeling method, (3)eye localization based on intensity information. Experimental results show that a correct eye detection rate of 98.9% can be achieved on 2408 FERET images with variations in lighting condition and facial expressions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.9
• /
• pp.2312-2325
• /
• 2013

In this paper, we propose a preprocessing method that it is to correct the distortion of text area in Korean signboard images as a preprocessing step to improve character recognition. Distorted perspective in recognizing of Korean signboard text may cause of the low recognition rate. The proposed method consists of four main steps and eight sub-steps: main step consists of potential vertical components detection, vertical components detection, text-boundary estimation and distortion correction. First, potential vertical line components detection consists of four steps, including edge detection for each connected component, pixel distance normalization in the edge, dominant-point detection in the edge and removal of horizontal components. Second, vertical line components detection is composed of removal of diagonal components and extraction of vertical line components. Third, the outline estimation step is composed of the left and right boundary line detection. Finally, distortion of the text image is corrected by bilinear transformation based on the estimated outline. We compared the changes in recognition rates of OCR before and after applying the proposed algorithm. The recognition rate of the distortion corrected signboard images is 29.63% and 21.9% higher at the character and the text unit than those of the original images.

• International Journal of Computer Science & Network Security
• /
• v.22 no.9
• /
• pp.175-182
• /
• 2022

In our daily life, we come across different types of information, for example in the format of multimedia and text. We all need different types of information for our common routines as watching/reading the news, listening to the radio, and watching different types of videos. However, sometimes we could run into problems when a certain type of information is required. For example, someone is listening to the radio and wants to listen to jazz, and unfortunately, all the radio channels play pop music mixed with advertisements. The listener gets stuck with pop music and gives up searching for jazz. So, the above example can be solved with an automatic audio classification system. Deep Learning (DL) models could make human life easy by using audio classifications, but it is expensive and difficult to deploy such models at edge devices like nano BLE sense raspberry pi, because these models require huge computational power like graphics processing unit (G.P.U), to solve the problem, we proposed DL model. In our proposed work, we had gone for a low complexity model for Audio Event Detection (AED), we extracted Mel-spectrograms of dimension 128×431×1 from audio signals and applied normalization. A total of 3 data augmentation methods were applied as follows: frequency masking, time masking, and mixup. In addition, we designed Convolutional Neural Network (CNN) with spatial dropout, batch normalization, and separable 2D inspired by VGGnet [1]. In addition, we reduced the model size by using model quantization of float16 to the trained model. Experiments were conducted on the updated dataset provided by the Detection and Classification of Acoustic Events and Scenes (DCASE) 2020 challenge. We confirm that our model achieved a val_loss of 0.33 and an accuracy of 90.34% within the 132.50KB model size.

• Journal of the Korea Computer Industry Society
• /
• v.9 no.1
• /
• pp.21-28
• /
• 2008

In this paper, the edge parameters were drawn from speaking image around lip and effective automatic lip reading system to recognize the Korean 'a/e/i/o/u' 5 owels were constructed using the parameter. Speaking images around lip were divided into $5{\times}5$ pane. In each pane the number of digital edge element using Sobel operator were evaluated. The observational error between samples was corrected by using normalization method and the normalized value is used for parameter In the experiment to convince the strength of parameter, 50 normal persons were sampled. The images of 10 persons were analyzed and the images of another 40 persons were experimented for recognition. 500 data are gathered and analyzed. Based on this analysis, the neural net system is constructed and the recognition experiments are performed for 400 data. The neural net system gave the best recognition result of 91.1%.

• Journal of Multimedia Information System
• /
• v.4 no.2
• /
• pp.79-82
• /
• 2017

Smoke and fire have different shapes and colours. This article suggests a fully connected system which is used two features using Adaboost algorithm for constructing a strong classifier as linear combination. We calculate the local histogram feature by gradient and bin, local binary pattern value, and projection vectors for each cell. According to the histogram magnitude, this paper applied adapted weighting value to improve the recognition rate. To preserve the local region and shape feature which has edge intensity, this paper processed the normalization sequence. For the extracted features, this paper Adaboost algorithm which makes strong classification to classify the objects. Our smoke detection system based on the proposed approach leads to higher detection accuracy than other system.

• Journal of Electrical Engineering and information Science
• /
• v.2 no.5
• /
• pp.72-78
• /
• 1997

This paper presents a fuzzy relaxation algorithm, which is based on the possibility and FAM instead of he probability and compatibility coefficients used in most of existing probabilistic relaxation algorithms, Because of eliminating stages for estimating of compatibility coefficients and normalization of the probability estimates, the proposed fuzzy relaxation algorithms increases the parallelism and has a simple iteration scheme. The construction of fuzzy relaxation scheme consists of the following three tasks: (1) definition of in/output linguistic variables, their term sets, and possibility. (2) Definition of FAM rule bases for relaxation using fuzzy compound relations. (3) Construction of the iteration scheme for calculating the new possibility estimate. Applications to region segmentation an ege detectiojn algorithms show that he proposed method can be used for not only reducing the image ambiguity and segmentation errors, but also enhancing the raw edge iteratively.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.6
• /
• pp.26-35
• /
• 2009

In this paper, we present morphology-based step region extraction and regularized iterative point-spread-function (PSF) estimation methods. The proposed PSF estimation method uses canny edge detector to extract the edge of the input image. We extract feasible vertical and horizontal edges using morphology analysis, such as the hit-or-miss transform. Given extracted edges we estimate the optimal step-response using flattening and normalization processes. The PSF is finally characterized by solving the equation which relates the optimal step response and the 2D isotropic PSF. We shows the restored image by the estimated PSF. The proposed algorithm can be applied a fully digital auto-focusing system without using mechanical focusing parts.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.55 no.4
• /
• pp.165-172
• /
• 2006

In this paper we present a technique for detecting cross-reference points that allows improving watermark detect-ability. In general, Harris detector is commonly used for finding salient points. Harris detector is a kind of combined corner and edge detector which is based on neighboring image data distribution, therefore it has some limitation to find accurate salient points after watermark embedding or any kinds of digital attacks. The new method proposed in this paper used not data distribution but geometrical structure of a normalized image in order to avoid pointing error caused by the distortion of image data. After normalization, we constructed pre-specified number of virtual lines from top to bottom and left to right, and several of cross points were selected by a random key. These selected points specify almost same positions with the accuracy more than that of Harris detector after digital attacks. These points were arranged by a random key, and blocks centered in these points were formed. A reference watermark is formed by a block and embedded in the next block. Because same alteration is applied to the watermark generated and embedded blocks. the detect-ability of watermark is improved even after digital attacks.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.11
• /
• pp.3761-3779
• /
• 2022

The performance of existing recognition algorithms for binary phase shift keying (BPSK) and quadrature phase shift keying (QPSK) signals degrade under conditions of low signal-to-noise ratios (SNR). Hence, a novel recognition algorithm based on features in the graph domain is proposed in this study. First, the power spectrum of the squared candidate signal is truncated by a rectangular window. Thereafter, the graph representation of the truncated spectrum is obtained via normalization, quantization, and edge construction. Based on the analysis of the connectivity difference of the graphs under different hypotheses, the sum of degree (SD) of the graphs is utilized as a discriminate feature to classify BPSK and QPSK signals. Moreover, we prove that the SD is a Schur-concave function with respect to the probability vector of the vertices (PVV). Extensive simulations confirm the effectiveness of the proposed algorithm, and its superiority to the listed model-driven-based (MDB) algorithms in terms of recognition performance under low SNRs and computational complexity. As it is confirmed that the proposed method reduces the computational complexity of existing graph-based algorithms, it can be applied in modulation recognition of radar or communication signals in real-time processing, and does not require any prior knowledge about the training sets, channel coefficients, or noise power.

• Journal of Korean Society for Geospatial Information Science
• /
• v.20 no.2
• /
• pp.47-54
• /
• 2012

Additional high resolution satellite images, other period or site, are essential for efficient city modeling and analysis. However, the same ground objects have a radiometric inconsistency in different satellite images and it debase the quality of image processing and analysis. Moreover, in an urban area, buildings, trees, bridges, and other artificial objects cause shadow effects, which lower the performance of relative radiometric normalization. Therefore, in this study, we exclude shadow areas and suggest the selective histogram matching methods for image based application without supplementary digital elevation model or geometric informations of sun and sensor. We extract the shadow objects first using adjacency informations with the building edge buffer and spatial and spectral attributes derived from the image segmentation. And, Outlier objects like a asphalt roads are removed. Finally, selective histogram matching is performed from the shadow masked multi-temporal Quickbird-2 images.