• Journal of Broadcast Engineering
• /
• v.11 no.1 s.30
• /
• pp.3-14
• /
• 2006

Music summarization means a technique which automatically generates the most importantand representative a part or parts ill music content. The techniques of music summarization have been studied with two categories according to summary characteristics. The first one is that the repeated part is provided as music summary and the second provides the combined segments which consist of segments with different characteristics as music summary in music content In this paper, we propose and evaluate two kinds of music summarization techniques. The algorithm using multi-level vector quantization which provides a repeated part as music summary gives fixed-length music summary is evaluated by overlapping ration between hand-made repeated parts and automatically generated summary. As results, the overlapping ratios of conventional methods are 42.2% and 47.4%, but that of proposed method with fixed-length summary is 67.1%. Optimal length music summary is evaluated by the portion of overlapping between summary and repeated part which is different length according to music content and the result shows that automatically-generated summary expresses more effective part than fixed-length summary with optimal length. The cluster-based algorithm using 2-D similarity matrix and k-means algorithm provides the combined segments as music summary. In order to evaluate this algorithm, we use MOS test consisting of two questions(How many similar segments are in summarized music? How many segments are included in same structure?) and the results show good performance.

• Journal of the Korean Society of Costume
• /
• v.22
• /
• pp.135-158
• /
• 1994

The fashion industry of today is closely re-lated with the information activity. This study investigates now they take advantage of the fashion information by analysing some related fields including fashion design. Furthormore this study identify the factor which makes the design activity distinctive and predict and fu-ture trend in design. The purposes of this study through analysis are : to help find the way to facilitate the fashion industry and to develop the fashion merchandising in order to activate the re-cession of manufacturers and to improve competitiveness in the world market and to contribute to the academic achievement in the study of fashion merchandising. In the experimental approach the tasks of research are as follows: Research 1: The application of fashion infor-mation is discussed in terms of process and level. Research 2: The differences of the fashion design activities are captured on the basis of the characteristics of manufactures and fashion specialists and fashion information ac-tivities. The factors are identified which pre-dict and discriminate the results of fashion de-sign activities. The questionaire and interview were conduc-ted among women's clothing manufactures in Seoul and their fashion specialists in charge of merchandise development. The methods of survey were designed on the basis of theories developed so far and on he advice from the social scientists and fashion specialists. the methods were corrected and complemented through the 1st and 2nd preliminary investigations before their application For data analysis cronbach's a coefficient fre-quency percentage average standard devi-ation pearson's correlation coefficient were calculated and t-test F-test(ANOVA) Duncan's multiple range test regression and Discriminant analysis were conducted. The results of analysis throughout the experimen-tal studies were as follows: 1. The analysis of fashion in formation (1) the present application of information a. The source of information : Both the foreign and the domestic information came mostly from fashion journals magazines newspapers and other periodicals, the main source of market information was collected from the survey of popular items in recent years the data about the life-style of customers and their tastes for wearing were widely used as the information about consumers. b. The applicaton of information : The most widely used information was about fabrics. The foreign information was prefered on the whole. The domestic problems pointed out in this investgation were that the domestic journals showed the lack of specialty and the special organizations for fashion informatino were in urgent need. (2) The research of fashion information a. The various processes and levels of in-formation activities: Among the process of collection analysis and distribution the collec-tion process showed a good acheivement whereas the distribution whereas the distribution didn't. In levels of the systematic activities the acceptance of support and the utility of supported instruments the first indicated a high degree whereas the second showed lowest. b. The correlationship among subvariables : There was a significant correlation between the collection and the analysis process. The systematic activities revealed close relation-ship with the analysis process and the accept-ance of support with the distribution process. The close correlation was found between the utility of supported instruments and the analy-sis process. 2. The analysis of the fashion design activities (1) No significant differences were found in the design activities when the characteristics of companies were compared only. (2) According to the characteristics of fashion specialists the one with age experi-ence and high income showed rather good achievement but no significant differences were captured among sex department in charge title academic background education in abroad and field experience (3) The fashion information activities were strongly correlated with the design achieve-ment : The analysis process and the system-atic activity level had a great influence on the design activities. (4) In order to examine which cha-racteristics in (1-3) made it possible to pre-dct and discriminate the achievement in de-sign activities the Discriminant analysis was carried out. The results were as follows: the fashion information activities showed the highest discriminant rate. Next came the in-come level experience and age in that order Those four variables discrimated 37 from 50 who showed great achievement in design area (74%) and 51 from 83 in low achievement group (81%) Thus the total discriminant rate was 77.5%.