• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.2
• /
• pp.380-389
• /
• 2005

The Goldschmidt iterative algorithm for a floating point divide calculates it by performing a fixed number of multiplications. In this paper, a variable latency Goldschmidt's divide algorithm is proposed, that performs multiplications a variable number of times until the error becomes smaller than a given value. To calculate a floating point divide '$\frac{N}{F}$', multifly '$T=\frac{1}{F}+e_t$' to the denominator and the nominator, then it becomes ’$\frac{TN}{TF}=\frac{N_0}{F_0}$'. And the algorithm repeats the following operations: ’$R_i=(2-e_r-F_i),\;N_{i+1}=N_i{\ast}R_i,\;F_{i+1}=F_i{\ast}R_i$, i$\in${0,1,...n-1}'. The bits to the right of p fractional bits in intermediate multiplication results are truncated, and this truncation error is less than ‘$e_r=2^{-p}$'. The value of p is 29 for the single precision floating point, and 59 for the double precision floating point. Let ’$F_i=1+e_i$', there is $F_{i+1}=1-e_{i+1},\;e_{i+1}',\;where\;e_{i+1}, If '$[F_i-1]<2^{\frac{-p+3}{2}}$ is true, ’$e_{i+1}<16e_r$' is less than the smallest number which is representable by floating point number. So, ‘$N_{i+1}$ is approximate to ‘$\frac{N}{F}$'. Since the number of multiplications performed by the proposed algorithm is dependent on the input values, the average number of multiplications per an operation is derived from many reciprocal tables ($T=\frac{1}{F}+e_t$) with varying sizes. 1'he superiority of this algorithm is proved by comparing this average number with the fixed number of multiplications of the conventional algorithm. Since the proposed algorithm only performs the multiplications until the error gets smaller than a given value, it can be used to improve the performance of a divider. Also, it can be used to construct optimized approximate reciprocal tables. The results of this paper can be applied to many areas that utilize floating point numbers, such as digital signal processing, computer graphics, multimedia, scientific computing, etc

• Cartoon and Animation Studies
• /
• s.6
• /
• pp.152-171
• /
• 2002

This study analyses the avatar service, which is recognized as an alternative strategy of animation industry. The research questions of this study are following: (1) How have the avatar services been developed and what are the present dominant types? (2) Which structural characteristics of e-business environment are needed for the success of avatar services? (3) What is the economic characteristics of avatar business model? To solve these research questions, the basic conditions and the structural characteristics of avatar services have been investigated. In the first place, two forms of avatar service are classified. One is the internet service site whole primary service is to provide chatting service based on avatar service. The other is the portal site in which many kinds of products and services are presented as bundles to meet the needs of internet users. So avatar service is one of bundles which those portal sites are providing with. In this study, the big five internet service sites are selected based on the profits they earned through the sales of avatar service. The result of analysis is that the pricing strategy of those big five sites is very different from those of traditional off-line markets. The pricing mechanism are based on the value which internet users endow with the avatar items, not based on the costs of making the products. Avatar is the representative informative goods. The informative goods have the original cost structures, constant fixed costs and zero marginal costs, so the providers of avatar services make much of the subjective values of consumers. The sayclub, which is the most successful avatar service site and earn the average sales of 3 billing won a month, takes the aggressive strategy of pricing avatar items at highest price in the industry. The avatar service providers which make lots of profits are planning of making differentiate the services, introducing well-known brand items and star-named items. Nevertheless, the fact that the members of the sayclub are not decreasing means that the network effect of the site is so strongly manifest. Moreover, the costs the members have paid for the avatar items are so big not as to switch from one site to the other site, it can be very costly. These switching costs are endemic in high-technology industries and digital contents industries. It can be so large that switching suppliers is virtually unthinkable, a situation known as 'lock-in'. When switching costs are substantial, competition can be intense to attract new customers, since, one they are locked in, they can be a substantial source of profit. The consumers of avatar items have switching costs if they subscribe for the new avatar service site. The switching costs can be subscription costs as well as the costs of giving up the items they already paid for. One common example of switching costs involves specialized supplies, as with inkjet printer cartridges. In this example, the switching cost is the purchase of a new printer. The market is competitive ex ante, but since cartridges are incompatible, it is monopolized dx post. So the providers of printer/cartridges set pricing printer so cheap and cartridges expensive. On the contrary, since the avatar service can be successful with the strong network effect, the providers of avatar services have to compete aggressively for new customers. So they allow the subscription at a low price(almost marginal cost) in the early market. The network effect can be maximized when the members are sufficiently growing. The providers which have the monopoly power with sufficient subscribers. begin to raise the prices over the lifetime of the product and make profits.

• Journal of Global Scholars of Marketing Science
• /
• v.19 no.2
• /
• pp.17-26
• /
• 2009

As the lifestyle of consumers changes and the need for various products increases, new products are being developed in the market. Each household owns various home appliances which are purchased through the choice of a decision maker. These appliances include not only large-sized products such as TV, refrigerator, and washing machine, but also small-sized products such as microwave oven and air cleaner. There exists latent correlation among possession of home appliances, even though they are purchased independently. The purpose of this research is to analyze the effect of demographic factors on the purchase and possession of each home appliances, and to derive some relationships among various appliances. To achieve this purpose, the present status on the possession of each home appliances are investigated through consumer survey data on the electric and energy product. And a multivariate probit(MVP) model is applied for the empirical analysis. From the estimation results, some appliances show a substitutive or complementary pattern as expected, while others which look apparently unrelated have correlation by co-incidence. This research has several advantages compared to previous literatures on home appliances. First, this research focuses on the various products which are purchased by each household, while previous researches such as Matsukawa and Ito(1998) and Yoon(2007) focus just on a particular product. Second, the methodology of this research can consider a choice process of each product and correlation among products simultaneously. Lastly, this research can analyze not only a substitutive or complementary relationship in the same category, but also the correlation among products in the different categories. As the data on the possession of home appliances in each household has a characteristic of multiple choice, not a single choice, a MVP model are used for the empirical analysis. A MVP model is derived from a random utility model, and has an advantage compared to a multinomial logit model in that correlation among error terms can be derive(Manchanda et al., 1999; Edwards and Allenby, 2003). It is assumed that the error term has a normal distribution with zero mean and variance-covariance matrix ${\Omega}$. Hence, the sign and value of correlation coefficients means the relationship between two alternatives(Manchanda et al., 1999). This research uses the data of 'TEMEP Household ICT/Energy Survey (THIES) 2008' which is conducted by Technology Management, Economics and Policy Program in Seoul National University. The empirical analysis of this research is accomplished in two steps. First, a MVP model with demographic variables is estimated to analyze the effect of the characteristics of household on the purchase of each home appliances. In this research, some variables such as education level, region, size of family, average income, type of house are considered. Second, a MVP model excluding demographic variables is estimated to analyze the correlation among each home appliances. According to the estimation results of variance-covariance matrix, each households tend to own some appliances such as washing machine-refrigerator-cleaner-microwave oven, and air conditioner-dish washer-washing machine and so on. On the other hand, several products such as analog braun tube TV-digital braun tube TV and desktop PC-portable PC show a substitutive pattern. Lastly, the correlation map of home appliances are derived using multi-dimensional scaling(MDS) method based on the result of variance-covariance matrix. This research can provide significant implications for the firm's marketing strategies such as bundling, pricing, display and so on. In addition, this research can provide significant information for the development of convergence products and related technologies. A convergence product can decrease its market uncertainty, if two products which consumers tend to purchase together are integrated into it. The results of this research are more meaningful because it is based on the possession status of each household through the survey data.

• Sleep Medicine and Psychophysiology
• /
• v.10 no.1
• /
• pp.52-60
• /
• 2003

Objectives: Stage 1 sleep provides important information regarding interpretation of nocturnal polysomnography, particularly sleep onset. It is a short transition period from wakeful consciousness to sleep. The lack of prominent sleep events characterizing stage 1 sleep is a major obstacle in automatic sleep stage scoring. In this study, utilization of simultaneous EEG and EOG processing and analyses to detect stage 1 sleep automatically were attempted. Methods: Relative powers of the alpha waves and the theta waves were calculated from spectral estimation. A relative power of alpha waves less than 50% or relative power of theta waves more than 23% was regarded as stage 1 sleep. SEM(slow eye movement) was defined as the duration of both-eye movement ranging from 1.5 to 4 seconds, and was also regarded as stage 1 sleep. If one of these three criteria was met, the epoch was regarded as stage 1 sleep. Results were compared to the manual rating results done by two polysomnography experts. Results: A total of 169 epochs were analyzed. The agreement rate for stage 1 sleep between automatic detection and manual scoring was 79.3% and Cohen’s Kappa was 0.586 (p<0.01). A significant portion (32%) of automatically detected stage 1 sleep included SEM. Conclusion: Generally, digitally-scored sleep staging shows accuracy up to 70%. Considering potential difficulty in stage 1 sleep scoring, accuracy of 79.3% in this study seems to be strong enough. Simultaneous analysis of EOG differentiates this study from previous ones which mainly depended on EEG analysis. The issue of close relationship between SEM and stage 1 sleep raised by Kinnari remains a valid one in this study.

• The Journal of Korean Academy of Prosthodontics
• /
• v.52 no.2
• /
• pp.105-112
• /
• 2014

Purpose: The aim of present investigation was to find out the influence of several times iatrogenic mobilization in the initial stage of implant installation on bone-implant osteointegration. Materials and methods: The experimental implants (3.75 mm in diameter, 8.0 mm in length) were made of commercially pure (Grade IV) titanium, and were treated with RBM ($MegaGen^{(R)}$: Ca-P) on lower 4.0 mm part. Only lower part of implant was inserted to bone and the implants were nonsubmerged. The 130 implants (two in each tibia) were inserted into the monocortical tibias of 33 rabbits which each weighed more than 3.5 kg (Female, New Zealand White). According to the removal torque interval, the groups were divided into 13 groups, group I (1 day), group II (1 day + 2 days), group III (1 day + 2 days + 3 days), group IV (1 day + 2 days + 3 days + 4 days), group V (2 days), group VI (2 days + 4 days), group VII (2 days + 4 days + 6 days), group VIII (2 days + 4 days + 6 days + 8 days), group IX (4 days), group X (4 days + 7 days), group XI (4 days + 7 days + 10 days), group XII (4 days + 7 days + 10 days + 14 days) and control group. In the control group, the removal torque was measured at 8 weeks with a digital torque gauge (Mark-10, USA). In the experimental groups, the removal torque was given once, twice, three times or four times by experiment design before the final removal torque and the value was measured each time. The implants were then screwed back to their original positions. All the experimental groups were given a final healing time of 8 weeks after placement, in which values were compared with the control groups and the 1st, 2nd, 3rd or 4th removal torque values in each experimental group. Results: In comparison of the final removal torque tests among experimental groups, the removal torque value of experimental groups except group XII were not statistically different that of control group. And the values of group I and II were significantly higher than the values of group VI, VIII, X, XI, and XII. In addition, the values of group III, IV, and V were significantly higher than group XI and XII. In comparison of the removal torque in the each experimental group, the final removal torque were significantly higher in all groups except group VIII, X, XI, and XII. Conclusion: If sufficient healing time was allowed, a few mobilization of fixture at the very early stage after the implant placement in the rabbits didn't interrupt the final bone to implant osseointegration.

• Progress in Medical Physics
• /
• v.15 no.1
• /
• pp.1-8
• /
• 2004

Patient dose verification is one of the most Important responsibilities of the physician in the treatment delivery of radiation therapy. For the task, it is necessary to use an accurate dosimeter that can verify the patient dose profile, and it is also necessary to determine the physical characteristics of beams used in intensity modulated radiation therapy (IMRT) The Beam Intensity Scanner (BInS) System is presented for the dosimetric verification of the two dimensional photon beam. The BInS has a scintillator, made of phosphor Terbium-doped Gadolinium Oxysulphide (Gd$_2$O$_2$S:Tb), to produce fluorescence from the irradiation of photon and electron beams. These fluoroscopic signals are collected and digitized by a digital video camera (DVC) and then processed by custom made software to express the relative dose profile in a 3 dimensional (3D) plot. As an application of the BInS, measurements related to IWRT are made and presented in this work. Using a static multileaf collimator (SMLC) technique, the intensity modulated beam (IMB) is delivered via a sequence of static portals made by controlled leaves. Thus, when static subfields are generated by a sequence of abutting portals, the penumbras and scattered photons of the delivered beams overlap in abutting field regions and this results in the creation of “hot spots”. Using the BInS, inter-step “hot spots” inherent in SMLC are measured and an empirical method to remove them is proposed. Another major MLC technique in IMRT, the dynamic multileaf collimator (DMLC) technique, has different characteristics from SMLC due to a different leaf operation mechanism during the irradiation of photon and electron beams. By using the BInS, the actual delivered doses by SMLC and DMLC techniques are measured and compared. Even if the planned dose to a target volume is equal in our experimental setting, the actual delivered dose by DMLC technique is measured to be larger by 14.8% than that by SMLC, and this is due to scattered photons and contaminant electrons at d$_{max}$.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.31 no.3
• /
• pp.375-382
• /
• 1986

This study was carried out to obtain the basic information for the clarification of spring growth habits mechanism of naked barleys. The isozyme patterns and activities of peroxidase in the young spike and leaf blade were analyzed during the differentiation and development of young spike. The characteristic differences between the normal and rosetted type were in c and g isozymes in young spike, and in i isozyme in the leaf blade. In the normal type, c and i isozymes disappeared at the stage of spi-kelet differentiation, g isozyme at the stage of flolet differentiation. But, in the rosetted type, those three isozymes remained in dark stained condition until the time of final sampling. Especially, those three isozymes were higher in the rosetted type than those in the normal type even at the stage of bract differentiation(BDS), just prior to the reproductive stage. The activities of peroxidase decreased slowly after BDS in the young spike and leaf blade in the normal type, While, in the rosetted type, increased linearly, and the degree of increasing was remarkable in the young spike. It was interesting that the degree of activities in young spike was higher in the rosetted type than that in the normal type even at BDS. From the above results, the remarkable differences of the isozyme patterns and activities at BDS between the normal and rosetted type were considered to be the physiological expression of the varieties concerned with the degree of spring growth habits.

• Journal of Chest Surgery
• /
• v.35 no.7
• /
• pp.530-534
• /
• 2002

Background: In aortic surgery, division and ligation of the left brachiocephalic vein(LBV) may improve exposure of the aortic arch but controversy continues about the safety of this division and whether a divided vein should be reanastomosed after arch replacement was completed. The safety of LBV division and the fate of the left subclavian venous drainage after LBV division were studied. Material and Method: From November 1998 to January 2001, planned division and ligation of the LBV on the mid-line after median sternotomy was peformed in 10 patients during the aortic surgery with the consideration of local anatomy and distal aortic anastomosis. Assessment for upper extremity edema and neurologic symptoms, measurement of venous pressure in the right atrium and left internal jugular vein, and digital subtraction venography(DSV) of the left arm were made postoperatively. Result: In 10 patients there was improvement in access to the aortic arch for procedures on the ascending aorta or aortic arch. The mean age of patients was 62 years(range 24 to 70). Follow-up ranged from 3 weeks to 13 months. One patient died because of mediastinitis from methicilline-resistant staphylococcus aureus strain. All patients had edema on the left upper extremity, but resolved by the postoperative day 4. No patient had any residual edema or difficulty in using the left upper extremity during the entire follow-up period. No patient had postoperative stroke. Pressure difference between the right atrium and left internal jugular vein was peaked on the immediate postoperative period(mean peak pressure difference = 25mmHg), but gradually decreased, then plated by the postoperative day 4. In all DSV studies left subclavian vein flowed across the midline through the inferior thyroid venous plexus. Conclusion: We conclude that division of LBV is safe and reanastomosis is not necessary if inferior thyroid vein, which is developed as a main bridge connecting the left subclavian vein with right venous system, is preserved.

• Progress in Medical Physics
• /
• v.17 no.4
• /
• pp.212-223
• /
• 2006

Positional accuracy of the on-board imager (OBI) isocenter with gantry rotation was presented in this paper. Three different type of automatic evaluation methods of discrepancies between therapeutic and OBI isocenter using digital image processing techniques as well as a procedure stated in the customer acceptance procedure (CAP) were applied to check OBI isocenter migration trends. Two kinds of kV x-ray image set obtained at OBI source angle of $0^{\circ},\;90^{\circ},\;180^{\circ},\;270^{\circ}$ and every $10^{\circ}$ and raw projection data for cone-beam CT reconstruction were used for each evaluation method. Efficiencies of the methods were also estimated. If a user needs to obtain an isocenter variation map with full gantry rotation, a method taking OBI image for every $10^{\circ}$ and fitting with 5th order polynomial was appropriate. However for a mere quality assurance (QA) purpose of OBI isocenter accuracy, it was adequate to use only four OBI Images taken at the OBI source angle of $0^{\circ},\;90^{\circ},\;180^{\circ}\;and\;270^{\circ}$. Maximal discrepancy was 0.44 mm which was observed between the OBI source angle of $90^{\circ}\;and\;180^{\circ}$ OBI isocenter accuracy was maintained below 0.5 mm for a year. Proposed QA program may be helpful to Implement a reasonable routine QA of the OBI isocenter accuracy without great efforts.

• The KIPS Transactions:PartA
• /
• v.12A no.5 s.95
• /
• pp.413-420
• /
• 2005

The Newton-Raphson iterative algorithm for finding a floating point reciprocal square mot calculates it by performing a fixed number of multiplications. In this paper, a variable latency Newton-Raphson's reciprocal square root algorithm is proposed that performs multiplications a variable number of times until the error becomes smaller than a given value. To find the rediprocal square root of a floating point number F, the algorithm repeats the following operations: '$X_{i+1}=\frac{{X_i}(3-e_r-{FX_i}^2)}{2}$, $i\in{0,1,2,{\ldots}n-1}$' with the initial value is '$X_0=\frac{1}{\sqrt{F}}{\pm}e_0$'. The bits to the right of p fractional bits in intermediate multiplication results are truncated and this truncation error is less than '$e_r=2^{-p}$'. The value of p is 28 for the single precision floating point, and 58 for the double precision floating point. Let '$X_i=\frac{1}{\sqrt{F}}{\pm}e_i$, there is '$X_{i+1}=\frac{1}{\sqrt{F}}-e_{i+1}$, where '$e_{i+1}{<}\frac{3{\sqrt{F}}{{e_i}^2}}{2}{\mp}\frac{{Fe_i}^3}{2}+2e_r$'. If '$|\frac{\sqrt{3-e_r-{FX_i}^2}}{2}-1|<2^{\frac{\sqrt{-p}{2}}}$' is true, '$e_{i+1}<8e_r$' is less than the smallest number which is representable by floating point number. So, $X_{i+1}$ is approximate to '$\frac{1}{\sqrt{F}}$. Since the number of multiplications performed by the proposed algorithm is dependent on the input values, the average number of multiplications Per an operation is derived from many reciprocal square root tables ($X_0=\frac{1}{\sqrt{F}}{\pm}e_0$) with varying sizes. The superiority of this algorithm is proved by comparing this average number with the fixed number of multiplications of the conventional algorithm. Since the proposed algorithm only performs the multiplications until the error gets smaller than a given value, it can be used to improve the performance of a reciprocal square root unit. Also, it can be used to construct optimized approximate reciprocal square root tables. The results of this paper can be applied to many areas that utilize floating point numbers, such as digital signal processing, computer graphics, multimedia, scientific computing, etc.