博碩士論文查詢聯邦系統首頁 聯絡信箱 關於我們 加入我們

本論文已被瀏覽 36 次, [ 造訪詳細資料與全文 ] 10 次,[ 回到前頁查詢結果 ] [ 重新搜尋 ]

不同遮蔽模式之中文字認知

Recognizing Chinese Characters with Different Occlusion Patterns

作者:廖卿枝
畢業學校:大同大學
出版單位:大同大學
核准日期:2015-09-15
類型:Electronic Thesis or Dissertation
權限:Copyright information available at source archive--ttu....

中文摘要

就增生認知負荷觀點,遮蔽局部文字是平面設計者為抓住使用者目光,刻意營造視覺效果而使用之設計手法。而此種手法之運用,向來僅憑設計者直覺判斷,缺乏從使用者認知之實證,提供明確設計之依據。本研究主要目的想探討使用者在這種條件之限制下,如何復原缺損筆畫,進而識別出文字,期望輔助設計者在文字設計時,能針對不同之中文字結構,快速決定最佳遮蔽比例以及方位,讓文字保有容易辨識與正確辨識之特徵,縮短設計者初衷與使用者認知之差距。
本研究分三階段;階段一目的為找出最適合之中文字遮蔽比例與斜置時最適合之動態角度,並以其結果作為後續製作實驗樣本之依據。研究方法以取材自高頻字、兩種文字結構之18個字,遮蔽每個文字每一方位寬度之1/9、2/9、3/9 三種比例,讓受測者以最適合之遮蔽比例排序並回答半結構式問題。階段二,主要目的想探討不同之文字結構、遮蔽位置以及教育背景,對於文字遮蔽辨識之速度與正確率的影響。以2×4×2三因子混合之實驗探討;兩個獨立變項為不同的教育背景(設計、非設計)與文字之遮蔽位置(上、下、左、右),一個相依變項為文字結構(分上中下與左中右三部分組成)。依變項為使用者的辨識速度與辨識正確率。階段三之實驗設計與刺激文字同於實驗二,但採遮雙邊模式,並針對不同一群人施測,目的除進一步探討實驗二之結果,也比較不同遮蔽模式下,具設計背景與否針對文字辨識速率與正確率之異同。
研究發現:中文字遮蔽以2/9之比例最適中;遮蔽文字之辨識特徵主要在於部首(或起始筆畫),其次為關鍵部件(偏旁)。針對此兩種文字結構,遮蔽單或雙邊,以遮右方(含右上、右下)可獲得辨識速度較快、正確率較高之閱讀效果,因為最不損及主要辨識特徵。不論遮蔽單或雙邊,「文字結構」與「遮蔽位置」兩個變項具關鍵性地交互影響;因此採取遮蔽文字設計前應先予以審慎判斷。當遮蔽雙邊時,設計背景者針對遮蔽任一方位之文字,辨識速度皆慢於非設計背者,可能原因為當文字筆畫遮蔽愈多時,具設計背景者越可能視之為文字圖形,影響辨識反應。
本研究結果有助於設計者使用該手法應用於海報、卡片、產品或包裝設計上,能快速判斷遮蔽文字之比例與方位,避開損及主要辨識特徵之錯誤。

英文摘要

From a germane cognitive load perspective, occluded a Chinese character is a design method used by graphic designers to create visual effects and capture people s attention. The design process is normally based on design intuition rather than user recognition, which provides a better reference for design and product criteria. This research therefore investigates how people recall the position of occlusion (when they see an occluded Chinese character) and accurately recognize that character. The research aims to help graphic designers to fast determine appropriate occlusion proportions as well as positions of Chinese characters with different structural types when they use occluded characters to create images in their designs. The research also aims to shorten the difference between design institution and user recognition by assist designers to keep the main recognition feature of Chinese characters and enable people to recognize them more quickly and accurately.
More specifically, this research is divided in three parts and carries out three recognition experiments. The first part of this research is to find the most appropriate occlusion proportion of the Chinese character, and explore appropriate rotated visual angles (diagonal laying) for occluding Chinese characters with different structural types. The result in the first part of the research is then used to design experiment samples (stimuli) for the other two experiments conducted in the research. In the first part of this research, an experiment is conducted based on 18 Chinese characters with two structural types and selected from most commonly used characters, and participants are asked to rank the appropriateness of occlusion proportion for the characters have been shown to them. Participants are also asked to answer to open questions in the experiment. The second part of this research is to explore the impact of composition structure, the position of occlusion, and educational background on the visual identification of Chinese characters in terms of recognition response time and accuracy. A 2 4 2 three-way ANOVA with interactions analysis is performed in this research. In the analysis, the one independent variable representing participant s educational background is a between-subjects, with two levels: design and non-design. The other independent variable is composition structure. The effect of composition structure is a within-subjects factorial design, with two levels: top-middle-bottom structure and left-middle-right structure. The position of occlusion is a between-subjects factorial design, with four levels: top, down, right, and left. There are two dependent variables: response time and response accuracy rate. Except for using the same 18 Chinese characters (stimuli), both the second and third experiments are conducted differently with regard to experiment participant and occlusion position (Chinese characters are occluded single-sided in the second experiment while it is double-sided for the third experiment). The third experiment aims to not only verify the result obtained from the second experiment but also test the effect of occlusion position on both recognition response time and accuracy for the participants with a design educational background.
The experiment result shows that the most appropriate occlusion proportion for a Chinese character is to cut off 2 cm from each side of the grid. For an occluded Chinese character, the primary recognition feature is the key radical (or initial strokes), and the secondary recognition feature is the key component (combination of strokes). For the two occlusion patterns (a character is occluded either single-sided or double-sided), it is best to occlude the right side (including the top-right and bottom-right sides) for a shorter recognition time and higher recognition accuracy since this has the least effect on the main recognition feature of a character. The position of occlusion influences interactively the structural composition regardless whether Chinese characters are occluded single side or double side; therefore, designers should judge both variables carefully before they apply the occlusion method in their designs. Moreover, experiment participants with an educational background in design are more likely to react slowly in recognizing a Chinese character with double-sided occlusions than the participants are not trained in design. This could due to the fact that Chinese characters with higher proportions of occlusion are more likely to be viewed as character pattern by participants educated in design.
This research result helps graphic designers to avoid making mistakes in occluding the main recognition feature of Chinese characters, and assists graphic designers to fast determine appropriate occlusion proportions and positions when they use Chinese characters to create images for posters, greeting cards, products and packages.

 

計畫贊助者: