• Title/Summary/Keyword: Individual gradient

Search Result 119, Processing Time 0.031 seconds

Automated Individual Tree Detection and Crown Delineation Using High Spatial Resolution RGB Aerial Imagery

  • Park, Tae-Jin;Lee, Jong-Yeol;Lee, Woo-Kyun;Kwak, Doo-Ahn;Kwak, Han-Bin;Lee, Sang-Chul
    • Korean Journal of Remote Sensing
    • /
    • v.27 no.6
    • /
    • pp.703-715
    • /
    • 2011
  • Forests have been considered one of the most important ecosystems on the earth, affecting the lives and environment. The sustainable forest management requires accurate and timely information of forest and tree parameters. Appropriately interpreted remotely sensed imagery can provide quantitative data for deriving forest information temporally and spatially. Especially, analysis of individual tree detection and crown delineation is significant issue, because individual trees are basic units for forest management. Individual trees in aerial imagery have reflectance characteristics according to tree species, crown shape and hierarchical status. This study suggested a method that identified individual trees and delineated crown boundaries through adopting gradient method algorithm to amplified greenness data using red and green band of aerial imagery. The amplification of specific band value improved possibility of detecting individual trees, and gradient method algorithm was performed to apply to identify individual tree tops. Additionally, tree crown boundaries were explored using spectral intensity pattern created by geometric characteristic of tree crown shape. Finally, accuracy of result derived from this method was evaluated by comparing with the reference data about individual tree location, number and crown boundary acquired by visual interpretation. The accuracy ($\hat{K}$) of suggested method to identify individual trees was 0.89 and adequate window size for delineating crown boundaries was $19{\times}19$ window size (maximum crown size: 9.4m) with accuracy ($\hat{K}$) at 0.80.

Managing Customer's Usage Behavior in a Multi-vendor Loyalty Program

  • Koo, Kay-Ryung;Woo, Won-Seok
    • Journal of Distribution Science
    • /
    • v.13 no.5
    • /
    • pp.5-14
    • /
    • 2015
  • Purpose - Loyalty programs enable retailers to maintain longer and better customer relationships. In successful services, customers actively use and value these programs. As the proximity to the goal (goal gradient) might signal active participation, this study empirically examines customer's goal gradient behavior in a multi-vendor loyalty program. We also consider the effect of customer's accrual diversity on goal gradients, which is a differentiating feature in a multi-vendor loyalty program, and is further examined. Research Design, Data, and Methodology - The data consists of6,646 OK Cashbag members' individual transaction records from 2006 to 2009. The goal gradient hypothesis was tested as an increase in both the speed and the amount of accumulated award points. Result - The findings suggest that the goal gradient is also observed in a multi-vendor loyalty program, occurring more strongly among members with high accrual diversity. Conclusions - The results indicate that customers with high accrual diversity attend strongly to goal gradients in multi-vendor loyalty programs; hence, it is important for such program managers to better inform members about affiliated partners.

A Study for Individual Identification by Discriminating the Finger Face Image (손가락 면 영상 판별에 의한 개인 식별 연구)

  • Kim, Hee-Sung;Bae, Byung-Kyu
    • Journal of Korea Multimedia Society
    • /
    • v.13 no.3
    • /
    • pp.378-391
    • /
    • 2010
  • In this paper, it is tested that an individual is able to be identified with finger face images and the results are presented. Special operators, FFG(Facet Function Gradient) masks by which the gradient of a facet function fit on a gray levels of image patches can be computed are used and a new procedure named F-algorithm is introduced to match the finger face images. The finger face image is divided into the equal subregions and each subregions are divided into equal patches with this algorithm. The FFG masks are used for convolution operation over each patch to produce scalar values. These values from a feature matrix, and the identity of fingers is determined by a norm of the elements of the feature matrices. The distribution of the norms shows conspicuous differences between the pairs of hand images of the same persons and the pairs of the different persons. This is a result to prove the ability of discrimination with the finger face image. An identification rate of 95.0% is obtained as a result of the test in which 500 hand images taken from 100 persons are processed through F-algorithm. It is affirmed that the finger face reveals to be such a good biometrics as other hand parts owing to the ability of discrimination and the identification rate.

Iris Recognition using Multi-Resolution Frequency Analysis and Levenberg-Marquardt Back-Propagation

  • Jeong Yu-Jeong;Choi Gwang-Mi
    • Journal of information and communication convergence engineering
    • /
    • v.2 no.3
    • /
    • pp.177-181
    • /
    • 2004
  • In this paper, we suggest an Iris recognition system with an excellent recognition rate and confidence as an alternative biometric recognition technique that solves the limit in an existing individual discrimination. For its implementation, we extracted coefficients feature values with the wavelet transformation mainly used in the signal processing, and we used neural network to see a recognition rate. However, Scale Conjugate Gradient of nonlinear optimum method mainly used in neural network is not suitable to solve the optimum problem for its slow velocity of convergence. So we intended to enhance the recognition rate by using Levenberg-Marquardt Back-propagation which supplements existing Scale Conjugate Gradient for an implementation of the iris recognition system. We improved convergence velocity, efficiency, and stability by changing properly the size according to both convergence rate of solution and variation rate of variable vector with the implementation of an applied algorithm.

Modeling the Hall-Petch Relation of Ni-Base Polycrystalline Superalloys Using Strain-Gradient Crystal Plasticity Finite Element Method (변형구배 결정소성 유한요소해석법을 이용한 니켈기 다결정 합금의 Hall-Petch 관계 모델링)

  • Choi, Yoon Suk;Cho, Kyung-Mox;Nam, Dae-Geun;Choi, Il-Dong
    • Korean Journal of Materials Research
    • /
    • v.25 no.2
    • /
    • pp.81-89
    • /
    • 2015
  • A strain-gradient crystal plasticity constitutive model was developed in order to predict the Hall-Petch behavior of a Ni-base polycrystalline superalloy. The constitutive model involves statistically stored dislocation and geometrically necessary dislocation densities, which were incorporated into the Bailey-Hirsch type flow stress equation with six strength interaction coefficients. A strain-gradient term (called slip-system lattice incompatibility) developed by Acharya was used to calculate the geometrically necessary dislocation density. The description of Kocks-Argon-Ashby type thermally activated strain rate was also used to represent the shear rate of an individual slip system. The constitutive model was implemented in a user material subroutine for crystal plasticity finite element method simulations. The grain size dependence of the flow stress (viz., the Hall-Petch behavior) was predicted for a Ni-base polycrystalline superalloy NIMONIC PE16. Simulation results showed that the present constitutive model fairly reasonably predicts 0.2%-offset yield stresses in a limited range of the grain size.

Individual differences in categorical perception: L1 English learners' L2 perception of Korean stops

  • Kong, Eun Jong
    • Phonetics and Speech Sciences
    • /
    • v.11 no.4
    • /
    • pp.63-70
    • /
    • 2019
  • This study investigated individual variability of L2 learners' categorical judgments of L2 stops by exploring English learners' perceptual processing of two acoustic cues (voice onset time [VOT] and f0) and working memory capacity as sources of variation. As prior research has reported that English speakers' greater use of the redundant cue f0 was responsible for gradient processing of native stops, we examined whether the same processing characteristics would be observed in L2 learners' perception of Korean stops (/t/-/th/). 22 English learners of L2 Korean with a range of L2 proficiency participated in a visual analogue scaling task and demonstrated variable manners of judging the L2 Korean stops: Some were more gradient than others in performing the task. Correlation analysis revealed that L2 learners' categorical responses were modestly related to individuals' utilizations of a primary cue for the stop contrast (VOT for L1 English stops and f0 for L2 Korean stops), and were also related to better working memory capacity. Together, the current experimental evidence demonstrates adult L2 learners' top-down processing of stop consonants where linguistic and cognitive resources are devoted to a process of determining abstract phonemic identity.

Adaptive Equalizations for Multipath Fading Channels in Mobile Communications Using the Individual Tap LMS Algorithm (개별탭 LMS 알고리듬을 이용한 이동통신 페이딩 채널의 적응 등화)

  • 김남용;강창언
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.16 no.8
    • /
    • pp.745-757
    • /
    • 1991
  • In this paper, the theoretical convergence property of the individual tap LMS algorithm for equaliz action if analyzed, and the pefomances over seeral multipath time vaing mobile radio channels are investigaed. The individual tap adjusting method of the tapped-elay line equalizer using LMS algorithm is proved to have Wiener optimum solution. It has more rapid convergence speed and lowe bit error rates than conventional Tdl LMS and gradient lattice equlizer in time invariant or time variant multipath channels.

  • PDF

Analysis of the excess MSE of the individual tap LMS(ITLMS) equalizer (개별탭 LMS 등화기의 초과 자승 평균오차 분석)

  • 김남용;이영조;강창언
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.17 no.3
    • /
    • pp.232-239
    • /
    • 1992
  • In this paper, an analytical expression of the excess MSE of ITLMS(Individual Tap LMS) which results from the random fluctuation of the distance between the actual and optimal coefficients due to noisy gradient estimates, is derived. The expression reveals that the excess MSE of the ITLMS algorithm is smaller that that of the LMS algorithm. The performance of the algorithm from the computer simulation for a fading time-dispersive channel verifies the theoretical results.

  • PDF

Determination of Individual Lanthanide Elements by Neutron Activation Using a New Comparator Technique

  • Lee, Chul;Yim, Yung-Chang;Chung, Koo-Soon
    • Nuclear Engineering and Technology
    • /
    • v.4 no.2
    • /
    • pp.83-89
    • /
    • 1972
  • The contents of the individual lanthanide elements are determined by neutron activation using a new comparator technique, which employs the short-lived radio-isotope of 56Mn as a neutron flux monitor. The total rare earths are separated as a group from the monazite sample before irradiation. After irradiation the rare earths are separated from each other by gradient elution with ammonium alpha-hydroxyisobutyrate using a cation-exchange column. The contents of 14 individual rare earths, from lutetium to lanthanum, are deter-mined.

  • PDF

Isolation of Monocytes with High Purity and Yield from Peripheral Blood Mononuclear Cells by Flotation Density Gradient Centrifugation (부유밀도구배 원심분리를 이용하여 말초혈액단핵구로부터 고순도 및 고수율의 단세포 분리방법)

  • Bae, Jae-Ho;Son, Cheol-Hun;Park, You-Soo;Kim, Dong-Wan;Kim, Sun-Hee;Kang, Chi-Dug
    • Journal of Life Science
    • /
    • v.19 no.6
    • /
    • pp.728-734
    • /
    • 2009
  • In this work, a simple, inexpensive and reproducible technique of flotation density gradient centrifugation was developed to isolate monocytes with high purity and yield from peripheral blood mononuclear cells (PBMC) using Histopaque solution, density and osmolarity of which were modified to 1.072 g/ml and 335 mOsm with phosphate-buffered saline (PBS) and sodium chloride (NaCl) solution, respectively. The average purity of monocytes was 74.75${\pm}$3.84%, with the individual purity ranging from 71.44% to 82.38%. The average yield of monocytes was 32.62${\pm}$11.16%, with the individual yields ranging from 21.02 to 53.63%. The monocytes isolated by floatation density gradient centrifugation could be successfully cultured into morphologically, phenotypically and functionally dendritic cells in vitro. In conclusion, the entire procedure seemed to be faster and more convenient, simple and cost-effective than other monocyte isolation methods, including plastic adherence and density gradient methods, and has the potential to be developed as a closed system for clinical scale generation of dendritic cells.