• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.4
• /
• pp.15-24
• /
• 2007

The recognition of postal address images is indispensable for the automatic sorting of postal envelopes. The process of the address image recognition is composed of three steps-address image preprocessing, character recognition, address interpretation. The extracted character images from the preprocessing step are forwarded to the character recognition step, in which multiple candidate characters with reliability scores are obtained for each character image extracted. aracters with reliability scores are obtained for each character image extracted. Utilizing those character candidates with scores, we obtain the final valid address for the input envelope image through the address interpretation step. The envelope sorting rate depends on the performance of all three steps, among which character recognition step could be said to be very important. The good character recognizer would be the one which could produce valid candidates with very reliable scores to help the address interpretation step go easy. In this paper, we propose the method of generating character candidates with reliable recognition scores. We utilize the existing MLP(multilayered perceptrons) neural network of the address recognition system in the current automatic postal envelope sorters, as the classifier for the each image from the preprocessing step. The MLP is well known to be one of the best classifiers in terms of processing speed and recognition rate. The false alarm problem, however, might be occurred in recognition results, which made the address interpretation hard. To make address interpretation easy and improve the envelope sorting rate, we propose promising methods to reestimate the recognition score (confidence) of the existing MLP classifier: the generation method of the statistical recognition properties of the classifier and the method of the combination of the MLP and the subspace classifier which roles as a reestimator of the confidence. To confirm the superiority of the proposed method, we have used the character images of the real postal envelopes from the sorters in the post office. The experimental results show that the proposed method produces high reliability in terms of error and rejection for individual characters and non-characters.

• Journal of Broadcast Engineering
• /
• v.23 no.4
• /
• pp.526-535
• /
• 2018

Many steganography algorithms have been studied, and steganalysis for detecting stego images which steganography is applied to has also been studied in parallel. Especially, in the case of the image steganalysis, the features such as ALE, SPAM, and SRMQ are extracted from the statistical characteristics of the image, and stego images are classified by learning the classifier using various machine learning algorithms. However, these studies did not consider the effect of image size, aspect ratio, or message-embedding rate, and thus the features might not function normally for images with conditions different from those used in the their studies. In this paper, we analyze the classification rate of the SPAM-based image stegnalysis against variety image sizes aspect ratios and message-embedding rates and verify its versatility.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.49 no.4
• /
• pp.23-30
• /
• 2012

This paper proposes an improved steganalysis method to detect the existence of secret message. Firstly, we magnify the small stego noise by multiplying the speckle noise to a given image and then we estimate the denoised image by using the soft thresholding method. Because the noises are not perfectly eliminated, some noises exist in the estimated cover image. If the given image is the cover image, then the remained noise will be very small, but if it is the stego image, the remained noise will be relatively large. The parent-child relationship in the wavelet domain will be slighty broken in the stego image. From this characteristic, we extract the joint statistical moments from the difference image between the given image and the denoised image. Additionally, four statistical moments are extracted from the denoised image for the proposed steganalysis method. All extracted features are used as the input of MLP(multilayer perceptron) classifier. Experimental results show that the proposed scheme outperforms previous methods in terms of detection rates and accuracy.

• International Journal of Fluid Machinery and Systems
• /
• v.8 no.4
• /
• pp.264-273
• /
• 2015

This presentation describes an experimental approach for the detection of cavitation in hydraulic machines by use of ultrasonic signal analysis. Instead of using the high frequency pulses (typically 1MHz) only for transit time measurement different other signal characteristics are extracted from the individual signals and its correlation function with reference signals in order to gain knowledge of the water conditions. As the pulse repetition rate is high (typically 100Hz), statistical parameters can be extracted of the signals. The idea is to find patterns in the parameters by a classifier that can distinguish between the different water states. This classification scheme has been applied to different cavitation sections: a sphere in a water flow in circular tube at the HSLU in Lucerne, a NACA profile in a cavitation tunnel and two Francis model test turbines all at LMH in Lausanne. From the signal raw data several statistical parameters in the time and frequency domain as well as from the correlation function with reference signals have been determined. As classifiers two methods were used: neural feed forward networks and decision trees. For both classification methods realizations with lowest complexity as possible are of special interest. It is shown that two to three signal characteristics, two from the signal itself and one from the correlation function are in many cases sufficient for the detection capability. The final goal is to combine these results with operating point, vibration, acoustic emission and dynamic pressure information such that a distinction between dangerous and not dangerous cavitation is possible.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.42 no.6
• /
• pp.75-82
• /
• 2005

The robust feature vector selection method for the multiple frontal face detection is proposed in this paper. The proposed feature vector for the training and classification are integrated by means, amplitude projections, and its 1D Harr wavelet of the input image. And the statistical modeling is performed both for face and nonface classes. Finally, the estimated probability density functions (PDFs) are applied for the detection of multiple frontal faces in the still image. The proposed method can handle multiple faces, partially occluded faces, and slightly posed-angle faces. And also the proposed method is very effective for low quality face images. Experimental results show that detection rate of the propose method is $98.3\%$ with three false detections on the testing data, SET3 which have 227 faces in 80 images.

• Journal of Electrical Engineering and Technology
• /
• v.11 no.6
• /
• pp.1825-1838
• /
• 2016

The detection of skin pigment is crucial in the diagnosis of skin diseases and in the evaluation of medical cosmetics and hairdressing. Accuracy in the detection is a basis for the prompt cure of skin diseases. This study presents a method to recognize and measure human skin pigment using Hemoglobin-Melanin (HM) coordinate. The proposed method extracts the skin area through a Gaussian skin-color model estimated from statistical analysis and decomposes the skin area into two pigments of hemoglobin and melanin using an Independent Component Analysis (ICA) algorithm. Then, we divide the two-dimensional (2D) HM coordinate into rectangular bins and compute the location histograms of hemoglobin and melanin for all the bins. We label the skin pigment of hemoglobin, melanin, and normal skin on all bins according to the Bayesian classifier. These bin-based HM projective histograms can quantify the skin pigment and compute the standard deviation on the total quantification of skin pigments surrounding normal skin. We tested our scheme using images taken under different illumination conditions. Several cosmetic coverings were used to test the performance of the proposed method. The experimental results show that the proposed method can detect skin pigments with more accuracy and evaluate cosmetic covering effects more effectively than conventional methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.6
• /
• pp.2806-2825
• /
• 2018

Unsupervised learning has shown good performance on image, video and audio classification tasks, and much progress has been made so far. It studies how systems can learn to represent particular input patterns in a way that reflects the statistical structure of the overall collection of input patterns. Many promising deep learning systems are commonly trained by the greedy layerwise unsupervised learning manner. The performance of these deep learning architectures benefits from the unsupervised learning ability to disentangling the abstractions and picking out the useful features. However, the existing unsupervised learning algorithms are often difficult to train partly because of the requirement of extensive hyperparameters. The tuning of these hyperparameters is a laborious task that requires expert knowledge, rules of thumb or extensive search. In this paper, we propose a simple and effective unsupervised feature learning algorithm for image classification, which exploits an explicit optimizing way for population and lifetime sparsity. Firstly, a sparse target matrix is built by the competitive rules. Then, the sparse features are optimized by means of minimizing the Euclidean norm ($L_2$) error between the sparse target and the competitive layer outputs. Finally, a classifier is trained using the obtained sparse features. Experimental results show that the proposed method achieves good performance for image classification, and provides discriminative features that generalize well.

• The Korean Journal of Ecology
• /
• v.19 no.5
• /
• pp.465-476
• /
• 1996

There are several environmental variables that may be influential to the spatial distribution of forest vegetation. To create a map of forest vegetation zone over southern part of Korean Peninsula, digital map layers were produced for each of environmental variables that include topography, geographic locations, and climate. In addition, an extensive set of field survey data was collected at relatively undisturbed forests and they were introduced into the GIS database with exact coordinates of survey sites. Preliminary statistical analysis on the survey data showed that the environmental variables were significantly different among the previously defined five forest vegetation zones. Classification of the six layers of digital map representing environmental variables was carried out by a supervised classifier using the training statistics from field survey data and by a clustering algorithm. Although the maps from two classifiers were somewhat different due to the classification procedure applied, they showed overall patterns of vertical and horizontal distribution of forest zones. considering the spatial contents of many ecological studies, GIS can be used as an important tool to manage and analyze spatial data. This study discusses more about the generation of digital map and the analysis procedure rather than the outcome map of forest vegetation zone.

• Journal of Information Technology Application
• /
• v.1
• /
• pp.173-211
• /
• 1999

The task of classification permeates all walks of life, from business and economics to science and public policy. In this context, nonlinear techniques from artificial intelligence have often proven to be more effective than the methods of classical statistics. The objective of knowledge discovery and data mining is to support decision making through the effective use of information. The automated approach to knowledge discovery is especially useful when dealing with large data sets or complex relationships. For many applications, automated software may find subtle patterns which escape the notice of manual analysis, or whose complexity exceeds the cognitive capabilities of humans. This paper explores the utility of a collaborative learning approach involving integrated models in the preprocessing and postprocessing stages. For instance, a genetic algorithm effects feature-weight optimization in a preprocessing module. Moreover, an inductive tree, artificial neural network (ANN), and k-nearest neighbor (kNN) techniques serve as postprocessing modules. More specifically, the postprocessors act as second0order classifiers which determine the best first-order classifier on a case-by-case basis. In addition to the second-order models, a voting scheme is investigated as a simple, but efficient, postprocessing model. The first-order models consist of statistical and machine learning models such as logistic regression (logit), multivariate discriminant analysis (MDA), ANN, and kNN. The genetic algorithm, inductive decision tree, and voting scheme act as kernel modules for collaborative learning. These ideas are explored against the background of a practical application relating to financial fraud management which exemplifies a binary classification problem.

• Journal of KIISE:Software and Applications
• /
• v.34 no.4
• /
• pp.377-388
• /
• 2007

In this paper, we propose a multi-modal biometric system based on face, iris and fingerprint recognition system. To effectively aggregate two systems, we use statistical distribution models based on matching values for genuine and impostor, respectively. And then, We performed reveal fusion algorithms including weighted summation, Support Vector Machine(SVM), Fisher discriminant analysis, Bayesian classifier. From the various experiments, we found that the performance of multi-modal biometric system was influenced with the normalization methods and classifiers.