• Journal of the Korean Institute of Intelligent Systems
• /
• v.23 no.2
• /
• pp.166-171
• /
• 2013

The method proposed in this paper can improve the performance of the Boosting algorithm in machine learning. The proposed Boundary AdaBoost algorithm can make up for the weak points of Normal binary classifier using threshold boundary concepts. The new proposed boundary can be located near the threshold of the binary classifier. The proposed algorithm improves classification in areas where Normal binary classifier is weak. Thus, the optimal boundary final classifier can decrease error rates classified with more reasonable features. Finally, this paper derives the new algorithm's optimal solution, and it demonstrates how classifier accuracy can be improved using the proposed Boundary AdaBoost in a simulation experiment of pedestrian detection using 10-fold cross validation.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.6
• /
• pp.1521-1530
• /
• 2014

The ROC analysis is considered for multiple class diagnosis. There exist many criteria to find optimal thresholds and measure the accuracy of diagnostic tests for k dimensional ROC analysis. In this paper, we proposed a diagnostic accuracy measure called the correct classification simple rate, which is defined as the summation of true rates for each classification distribution and expressed as a function of summation of sequential true rates for two consecutive distributions. This measure does not weight accuracy across categories by the category prevalence and is comparable across populations for multiple class diagnosis. It is found that this accuracy measure does not only have a relationship with Kolmogorov - Smirnov statistics, but also can be represented as a linear function of some optimal threshold criteria. With these facts, the suggested measure could be applied to test for comparing multiple distributions.

• Journal of the Korea Society of Computer and Information
• /
• v.6 no.4
• /
• pp.1-5
• /
• 2001

A new approach for determining global threshold value for image binarization is proposed in this paper. In the proposed algorithm, bit-plane information which involve the shapes of original image is used for dividing image into two parts object and background, and then compared each average values. Optimal threshold value are selected in center of two averages. Proposed method is relatively simple but robust and achieved good results in continuous tone images and document image.

• Proceedings of the KIEE Conference
• /
• 1987.07b
• /
• pp.1371-1374
• /
• 1987

This paper presents a new automatic multithreshold selection method which is based on the threshold selection method proposed by Otsu. This method can overcome some of limitations of the Otsu's method. An optimal threshold is selected by the new criterion so as to maximize the separability in all subregions. To get multiple thresholds, the procedure may be recursively applied to the resultant classes which are determined by the proposed evaluation measure.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.73-76
• /
• 1999

In this paper, we propose a new design of turbo decoder. It contains the simple additional unit which automatically decides the number of the iteration by detecting of the reliability value as threshold value. We investigate the relationship between the reliability value and the number of the iteration. We find the optimal threshold value without noticeable loss in performance. As a results of the simulation, it reduces the average number of the iteration compared with the conventional turbo decoder.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.37 no.1
• /
• pp.48-55
• /
• 2024

The changes in threshold voltage and DIBL were investigated for changes in remanent polarization P_r and coercive field E_c, which determine the characteristics of the P-E hysteresis curve of ferroelectric in NCFET (negative capacitance FET). The threshold voltage and DIBL (drain-induced barrier lowering) were observed for a junctionless double gate MOSFET using a gate oxide structure of MFMIS (metal-ferroelectric-metal-insulator-semiconductor). To obtain the threshold voltage, series-type potential distribution and second derivative method were used. As a result, it can be seen that the threshold voltage increases when P_r decreases and E_c increases, and the threshold voltage is also maintained constant when the P_r/E_c is constant. However, as the drain voltage increases, the threshold voltage changes significantly according to P_r/E_c, so the DIBL greatly changes for P_r/E_c. In other words, when P_r/E_c=15 pF/cm, DIBL showed a negative value regardless of the channel length under the conditions of ferroelectric thickness of 10 nm and SiO₂ thickness of 1 nm. The DIBL value was in the negative or positive range for the channel length when the P_r/E_c is 25 pF/cm or more under the same conditions, so the condition of DIBL=0 could be obtained. As such, the optimal condition to reduce short channel effects can be obtained since the threshold voltage and DIBL can be adjusted according to the device dimension of NCFET and the P_r and E_c of ferroelectric.

• Journal of Korean Institute of Industrial Engineers
• /
• v.34 no.4
• /
• pp.398-407
• /
• 2008

We consider a product recovery system that a single product is stocked in order to meet a demand from customers who may return products after usage. This paper addresses a problem of when to release a procurement process to replenish serviceable inventory and how many new products to procure. The structure of the optimal procurement policy is examined and numerically identified as a monotonic threshold curve. A numerical procedure is presented to jointly find the optimal procurement order size, optimal procurement policy, and optimal discounted profit. Sensitivity analysis also indicates that these optimal performance measurements have monotonic properties with respect to system parameters.

• Journal of the Korea Computer Industry Society
• /
• v.2 no.3
• /
• pp.389-396
• /
• 2001

In detection of a scene change of the moving pictures which has massive information capacity, the temporal sampling method has a faster searching speed and lower missing scene change detection than the sequential searching method for the whole moving pictures, yet employed searching algorithm and detection interval greatly affect missing frame and searching precision. In this study, the whole moving pictures were primarily retrieved threshold by the temporal difference of histogram scene change detection method. We suggested a dynamic threshold algorithm using cut detection interval and derived an equation formula to determine optimal primary retrieval threshold which can cut detection interval computation. Experimental results show that the proposed dynamic threshold algorithm using cut detection interval method works up about 30 percent in precision of performance than the sequential searching method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.7
• /
• pp.3524-3542
• /
• 2017

In a single-source and multi-relay amplify-forward (AF) cooperative network, the outage probability and the power allocation are two key factors to influence the performance of an entire system. In this paper, an optimized AF relay selection by an exclusive method and near optimal power allocation (NOPA) is proposed for both good outage probability and power efficiency. Given the same power at the source and the relay nodes, a threshold for selecting the relay nodes is deduced and employed to minimize the average outage probability. It mainly excludes the relay nodes with much higher thresholds over the aforementioned threshold and thus the remainders of the relay nodes participate in cooperative forwarding efficiently. So the proposed scheme can improve the utility of the resources in the cooperative multi-relay system, as well as reduce the computational complexity. In addition, based on the proposed scheme, a NOPA is also suggested to approach sub-optimal power efficiency with low complexity. Simulation results show that the proposed scheme obtains about 2.1dB and 5.8dB performance gain at outage probability of $10^{-4}$, when compared with the all-relay-forward (6 participated relays) and the single-relay-forward schemes. Furthermore, it obtains the minimum outage probability among all selective relay schemes with the same number of the relays. Meanwhile, it approaches closely to the optimal exhaustive scheme, thus reduce much complexity. Moreover, the proposed NOPA scheme achieves better outage probability than those of the equal power allocation schemes. Therefore, the proposed scheme can obtain good outage probability, low computational complexity and high power efficiency, which makes it pragmatic efficiently in the single-source and multi-relay AF based cooperative networks.

• Journal of KIISE:Databases
• /
• v.30 no.4
• /
• pp.362-371
• /
• 2003

In this paper, we propose an optimal interest point extraction method to support shape-base image classification and indexing for image database by applying a dynamic threshold that reflects the characteristics of the shape contour. The threshold is determined dynamically by comparing the contour length ratio of the original shape and the approximated polygon while the algorithm is running. Because our algorithm considers the characteristics of the shape contour, it can minimize the number of interest points. For n points of the contour, the proposed algorithm has O(nlogn) computational cost on an average to extract the number of m optimal interest points. Experiments were performed on the 70 synthetic shapes of 7 different contour types and 1100 fish shapes. It shows the average optimization ratio up to 0.92 and has 14% improvement, compared to the fixed threshold method. The shape features extracted from our proposed method can be used for shape-based image classification, indexing, and similarity search via normalization.