• The KIPS Transactions:PartD
• /
• v.8D no.3
• /
• pp.273-284
• /
• 2001

A systematic approach for interpreting Korean addresses based on postal code is presented in this paper. The implementation is focused on producing the final delivery point code from various types of address recognized. There are two stages in the address interpretation : 1) agreement verification between the recognized postal code and upper part of the address and 2) analysis of lower part of the address. In the agreement verification procedure, the recognized postal code is used as the key to the address dictionary and each of the retrieved addresses is compared with the words in the recognized address. As the result, the boundary between the upper part and the lower part is located. The confusion matrix, which is introduced to correct possible mis-recognized characters, is applied to improve the performance of the process. In the procedure for interpreting the lower part address, a delivery code is assigned using the house number and/or the building name. Several rules for the interpretation have been developed based on the real addresses collected. Experiments have been performed to evaluate the proposed approach using addresses collected from Kwangju and Pusan areas.

• Journal of Digital Contents Society
• /
• v.19 no.1
• /
• pp.141-147
• /
• 2018

This study analyzed the early retired employees who stayed in company no longer than 3 years based on a certain company's personality evaluation result data. The predicted model was analyzed by dividing into two categories; the manufacture group and the R&D group. Independent variables were selected according to the stepwise method. A logistic regression model was selected as a prediction model among various supervised learning methods, and trained through cross-validation to prevent over-fitting or under-fitting. The accuracy of the two groups were confirmed by the confusion matrix. The most influential factor for early retirement in the manufacture group was revealed as "immersion," and for the R&D group appeared as "antisocial." In the past, people concentrated on collecting data by questionnaire and identifying factors that are highly related to the retirement, but this study suggests a sustainable early retirement prediction model in the future by analyzing the tangible outcome of the recruitment process.

• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.225-238
• /
• 2021

The ROC, TOC, and TROC curves, which are visually descriptive methods of exploring the performance of the binary classification model, are implemented with TP, TN, FP, FN which consist of the confusion matrix, as well as their ratios TPR, TNR, FPR, FNR. In this study, we consider two types odds and then propose an odds curve representing these odds. And show the relationship between the odds curve and ROC curve. Based on the odds curve, we propose not only two statistics that measure the discriminant power of the odds curve but also the criteria for validation ratings of the odds curve. According to the shape of the odds curves, two classification distributions can be estimated and a criterion for validation ratings can be determined. The odds curve can be meaningfully used like other visual methods, and two kinds of measures for the discriminant power can be also applied together as an alternative criterion.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.61-69
• /
• 2022

As the delivery industry grew around the restaurant industry along with the COVID-19 situation, the number of delivery workers increased significantly. Along with that, new forms of delivery using personal mobility (PM) also emerged and two-wheeled or PM-related accidents are steadily increasing. This study manufactures a PM's driving analysis device to establish a safe delivery monitoring environment. This system was constructed to process data collected from the driving analysis device and through a cloud server, which would recognize and record special situations (acceleration/deceleration, speed bump) that could occur during the PM's driving situation. As a result, the angular speed, acceleration, and geomagnetic values collected from the IMU in the device were able to determine whether to drive, drive on the sidewalk, and drive on the speed bump. This technology was able to achieve approximately 1600 times more driving information storage efficiency than conventional image-based recording devices.

• Journal of Software Assessment and Valuation
• /
• v.17 no.2
• /
• pp.1-10
• /
• 2021

The ability to recognize human emotions by computer vision is a very important task, with many potential applications. Therefore the demand for emotion recognition using not only RGB images but also thermal images is increasing. Compared to RGB images, thermal images has the advantage of being less affected by lighting conditions but require a more sophisticated recognition method with low-resolution sources. In this paper, we propose a Divide and Conquer-based CNN training strategy to improve the performance of facial thermal image-based emotion recognition. The proposed method first trains to classify difficult-to-classify similar emotion classes into the same class group by confusion matrix analysis and then divides and solves the problem so that the emotion group classified into the same class group is recognized again as actual emotions. In experiments, the proposed method has improved accuracy in all the tests than when recognizing all the presented emotions with a single CNN model.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.3
• /
• pp.554-559
• /
• 2009

This paper proposes a new method of weighted template matching fur machine-printed numeral recognition. The proposed weighted template matching, which emphasizes the feature of a pattern using adaptive Hamming distance on local feature areas, improves the recognition rate while template matching processes an input image as one global feature. The experiment compares confusion matrices of the template matching, error back propagation neural network classifier, and the proposed weighted template matching respectively. The result shows that the proposed method improves fairly the recognition rate of the machine-printed numerals.

• 한국공간정보시스템학회:학술대회논문집
• /
• 2007.06a
• /
• pp.127-132
• /
• 2007

위성영상 기술의 발달과 고해상도 위성영상의 해상도 규제가 완화됨에 따라 건물의 높이 정보를 획득하는데 있어 고해상도 위성영상의 그림자 정보를 이용하는 연구들이 활발히 수행되어지고 있다. 그림자 정보를 이용하여 건물 높이 정보를 획득하는 연구의 정확도를 높이기 위해서는 정확한 건물의 그림자 탐지가 선행되어야 한다. 따라서 본 논문에서는 단영상을 이용한 그림자 탐지기법인 임계값법(Thresholding), 영상분류법, 영역확장법(Region Growing)을 건물의 그림자 탐지에 적용하여 각 기법의 장단점과 정확도를 평가하였다. 영상에서 수동으로 건물의 그림자를 디지타이징한 참조 자료와 기법들을 적용하여 탐지한 결과 영상을 시각적으로 비교하였고, 오차행렬(Confusion Matrix)을 이용한 전체정확도(Accuracy), F-measure, AOR(Area Overlap Ratio)을 이용하여 정량적인 정확도평가를 수행하였다. 실험결과 영역확장법을 적용한 경우 시각적 정량적으로 가장 높은 정확도를 보였으며, 영상분류법을 적용한 경우 시각적으로는 임계값을 적용한 경우보다 좋은 결과를 보였으나 정량적으로는 가장 낮은 정확도를 보였다.

• Proceedings of the KSME Conference
• /
• 2007.05a
• /
• pp.849-854
• /
• 2007

This paper describes design and tuning of a full-range Adaptive Cruise Control (ACC) with collision avoidance. The control scheme is designed to control the vehicle so that it would feel natural to the human driver and passengers during normal safe driving situations and to avoid rear-end collision in vehicle following situations. In this study, driving situations are determined using a non-dimensional warning index and time-to-collision (TTC). A confusion matrix method based on natural driving data sets was used to tune control parameters in the proposed ACC System. An ECU-Brake Hardware-in-the-loop Simulation (HiLS) was developed and used for an evaluation of ACC System. The ECU-Brake HiLS results for alternative driving situation are compared to manual driving data measured on actual traffic way. The ACC/CA control logic implemented in an ECU was tested using the ECU-Brake HiLS in a real vehicle environment.

• International journal of advanced smart convergence
• /
• v.6 no.3
• /
• pp.17-21
• /
• 2017

In this paper, the service execution accuracy was compared by ontology based rule inference method and machine learning method, and the amount of data at the point when the service execution accuracy of the machine learning method becomes equal to the service execution accuracy of the rule inference was found. The rule inference, which measures service execution accuracy and service execution accuracy using accumulated data and pattern matching on service results. And then machine learning method measures service execution accuracy using cross validation data. After creating a confusion matrix and measuring the accuracy of each service execution, the inference algorithm can be selected from the results.

• Proceedings of the KSRS Conference
• /
• 1998.09a
• /
• pp.213-217
• /
• 1998

This study overviews existing evaluation methods of classification accuracy using confusion matrix proposed by Cohen in 1960's, and proposes ISDd(Index of Spatial Distribution by distance) and ISDs(Index of Spatial Distribution by scatteredness) for the evaluation of spatial distribution of satellite image classification errors, which has not been tried yet. Index of spatial distribution offers the basis of decision on adoption/rejection of classification results at sub-image level by evaluation of distribution, such as status of local aggregation of misclassified pixels. So, users can understand the spatial distribution of misclassified pixels and, can have the basis of judgement of suitability and reliability of classification results.