WWW.THESIS.XLIBX.INFO
FREE ELECTRONIC LIBRARY - Thesis, documentation, books
 
<< HOME
CONTACTS



Pages:     | 1 |   ...   | 7 | 8 || 10 |

«A dissertation submitted to the Department Of Computer Science, Faculty of Science at the University Of Cape Town in partial fulfilment of the ...»

-- [ Page 9 ] --

Figure 5-8: The qualitative results for Question 5 Figure 5-9: The qualitative results for Question 9 “I had no problem seeing the facial expressions in this “It was difficult to follow the facial expressions in this video.” for each of the three frame rates and two video.” for each of the three frame rates and two resolutions. With a significance level of 0.330 (p =.330) resolutions. With a significance level of 0.792 (p =.792) there was no statistically significant difference in the there was no statistically significant difference in the average participant rating for each of the video clips. average participant rating for each of the video clips.

Figure 5-10: The qualitative results for Question 4 Figure 5-11: The qualitative results for Question 7 “The movement was blurry.” for each of the three frame “The movement was clear.” for each of the three frame rates and two resolutions. With a significance level of rates and two resolutions. With a significance level of 0.552 (p =.552) there was no statistically significant 0.416 (p =.416) there was no statistically significant difference in the average participant rating for each of the difference in the average participant rating for each of the video clips. video clips.

Figure 5-12: The qualitative results for Question 16 Figure 5-13: The qualitative results for Question 17 “I knew all the signs used in this video.” for each of the “Some signs used in this video were unknown to me.” for three frame rates and two resolutions. With a significance each of the three frame rates and two resolutions. With a level of 0.651 (p =.651) there was no statistically significance level of 0.670 (p =.670) there was no significant difference in the average participant rating for statistically significant difference in the average each of the video clips. participant rating for each of the video clips.

Figure 5-14: The qualitative results for Question 6 Figure 5-15: The qualitative results for Question 15 “The video was the right speed.” for each of the three “The video was too slow.” for each of the three frame rates frame rates and two resolutions. With a significance level and two resolutions. With a significance level of 0.133 (p of 0.913 (p =.913) there was no statistically significant =.133) there was no statistically significant difference in difference in the average participant rating for each of the the average participant rating for each of the video clips.

video clips.

–  –  –

Figure 5-17: The qualitative results for Question 11 Figure 5-18: The qualitative results for Question 13 “I had difficulty to understand what was said in this “It was easy to understand what was said in this video.” video.” for each of the three frame rates and two for each of the three frame rates and two resolutions. With resolutions. With a significance level of 0.540 (p =.540) a significance level of 0.502 (p =.502) there was no there was no statistically significant difference in the statistically significant difference in the average average participant rating for each of the video clips. participant rating for each of the video clips.

Figure 5-19: The qualitative results for Question 2 “I am sure of my answer to Question 1.” for each of the three frame rates and two resolutions. With a significance level of 0.227 (p =.227) there was no statistically significant difference in the average participant rating for each of the video clips.

Figure 5-20: Estimated marginal means across all questions.

The estimated marginal means of the participant responses across all questions for each of the two frame rates and two resolutions. The y-axis is mean total score. On the x-axis is a particular video resolution, with each colour representing a particular video frame rates. With a significance level of 0.990 (p =.990) there was no statistically significant difference in the average participant rating for each of the video clips.

No clear preference by the participants was found for any particular combination of frame rate and resolution, as can be seen from the estimated marginal means of the participant responses across all questions for each of the two frame rates and two resolutions, and is confirmed by the ANOVA analysis.

Despite getting feedback about possible Sign Language dialect problems and attempting to only use signed phrases without dialect problems, one of the four phrases (“He is a short man”) still was not caught before the experiment.

Where the two pilot studies were done with the help of DCCT staff members, all fully literate, the final experiment only included six DCCT staff members, with the remainder of varying literacy level. This could have impacted on the quality of the captured reponses.

No specific questions pertaining to the video intelligibility and quality gave any stand out problems. Of the four sign language video clips the only one that had problems contents wise was the “He is a short man” clip, and after pointing out the problem sign, participants continued to evaluate the video clip as per the questionnaire.

All lessons learnt through the pilot studies were applied, yet there were no frame rate and resolution combination that were clearly judged inadequate or below par.

6 Conclusion

6.1 Conclusion This dissertation studied the effects of reducing the frame rate and resolution of SASL video played back and viewed on a lower-end, more affordable cell phone. The subjective intelligibility of Sign Language videos at the different frame rates and resolutions were evaluated through user studies with members of the South African Deaf community.





When this research started, looking at similar research it seemed a relatively simple question:

What is the lowest video resolution and frame rate that would provide intelligible South African Sign Language video on a cell phone? But after two pilot studies and a final intelligibility study, what has become abundantly clear was that measuring the intelligibility of Sign Language video is a multifaceted problem, providing many obstacles, amongst others the difficulty with written language by the participants, making the use of written questionnaires problematic. In addition to this is the fact that SASL has different dialects, meaning a simple sign for one Deaf person could be an unknown sign to another. Each of these facets adds a layer of possible miscommunication and misunderstanding between the researcher and the Deaf participant that could impact on the evaluation of intelligibility.

Based on the results there does not seem to be a preferred frame rate or a clear drop in subjective intelligibility at low frame rate or low resolutions. The frame rate can be reduced to 10 frames per second, with the resolution reduced to 176 x 144 pixels while still providing intelligible SASL reproduction when viewed on a handheld cell phone, and being acceptable and comfortable for day to day use.

The final conclusion of this study is that, using long questionnaires and simple signed phrases, there is no clearly discernible difference between Deaf participants’ opinion of the intelligibility of sign language video clips based purely on frame rate and resolution.

6.2 Limitations Further work is needed in the subjective evaluation of the intelligibility of Sign Language video on a cell phone. While Nakazono et al. [19] used sign language video sequences of similar length as this research, about 7 to 8 seconds; other subjective assessments used longer video sequences, such as Cavender et al. [4] using clips with durations from 0:58 to 2:57 minutes and Ciaramello et al. [7] with video sequences ranging from 7.2 seconds to 150.9 seconds.

Short signed phrases might be good for ease of testing, but from the results of this research does not seem to be appropriate for the evaluation of intelligibility of Sign Language video sequences. A normal conversation consists of multiple longer sentences and provides the opportunity to clear up a missed or unknown word or sign by continuing to listen and possibly collect the missing information from its context. If the listener were to fail in this, it is simple to ask the speaker to repeat or explain the missed or unfamiliar word or sign. This conversational context is missing in the short phrases used in this research. A better approach might be to use a longer video clip showing a signer signing a short story about what happened that morning on the way to work, a part of life familiar to most people.

Looking at the questionnaires used in the related research they consisted of similar but fewer and simpler questions. The subjective questionnaire used by Ciaramello et al. [7] consisted of 12 videos in total with a four-question, multiple-choice survey focussing how difficult it was to understand what was said in the video. In the case of Cherniavsky et al. [5] the participants used the phone to hold a sign language conversation at different video settings with a five question subjective measurement after a five minute conversation. A five level scale was used to evaluate the video intelligibility.

In this research paper a similar questioning technique was followed with the participants being asked to subjectively evaluate sign language videos of differing video quality on a five level scale (except in experiment two where a binary answering technique was used). But in hindsight the questionnaire became too in-depth too early. Increasing the number of questions and the complexity of the questions increased the burden on the Deaf participants, introducing the additional complication of user fatigue, especially on the last experiment where English literacy was even less prevalent.

Looking at participant numbers in the related research these varied from 11 participants used by Ciaramello et al. [7] to 15 participants recruited by Cherniavsky et al. [5], compared to five, six and twenty four participants respectively in the three experiments in this research paper.

In the findings of both Ciaramello et al. [7] and Nakazono et al. [19] a clear progression was found from low intelligibility at low resolutions and frame rates to increased intelligibility scores at the higher quality video clips. This was not clearly evident in the results of this study. A clearer comparison between the results of this study and the related studies would have been possible if similar questions (in number and kind) were used with a larger number of Deaf users.

The literacy level of the participants and the grammatical differences between Sign Language and written language has an impact on the use of questionnaires in this research. The use of fewer, simpler questions could ease the execution of the experiments as well as improve the usefulness of the results.

To improve the results and attain statistical significance the sample size needs to be increased, the questionnaire shortened and simplified, and the experiments should make use of longer video clips. All related work had consistently longer video clips, more closely simulating a conversational use of the phone. By shortening and simplifying the questionnaire participant fatigue would be decreased and a more honest evaluation of the intelligibility would be captured.

6.3 Future work Open questions are: Are simple, short signed phrases evaluated through a written questionnaire a valid test of intelligibility and Sign Language communications over video? What is the impact of Sign Language dialects and sign execution by the signer on the evaluation of the clip? Can the impact of differences in Sign Language usage be negated or quantified during intelligibility evaluations?

These results indicate that we do not need to use the latest smart phone with high resolution video capabilities to provide the Deaf community the opportunity to converse in their first language wherever and whenever they want, bringing us closer to providing Deaf people affordable, full access to the mobile telecommunications network.

This research looked at the frame rate and resolution requirements using Sign Language video clips pre-recorded in controlled conditions, with good lighting and an even background. Further research is needed to confirm that the findings of this research hold up under non-ideal conditions, especially when the video to be viewed is recorded using the built-in camera of the cell phone. In addition there is the question of how much bandwidth, and thus cost, is truly saved by the drop in frame rate and resolution. Taking into consideration bandwidth cost as well as phone cost do these findings lead to widely affordable mobile video communication for the South African Deaf community?

Another avenue for future research would be to remove the written questionnaire and conduct the experiment completely in SASL instead of text. Would minimising the impact of participant literacy level and keeping the whole experimental setup as natural as possible for the participants improve the significance of the findings?

Bibliography

1. Ascher, S., Pincus, E. The Filmmaker’s Handbook – A comprehensive guide for the digital age. Third Edition (2007), Plume, New York, USA.

2. AV Foundation Programming Guide. 2010. [online]. [Accessed 2 December 2010]. Available

from World Wide Web:

https://developer.apple.com/library/ios/#documentation/AudioVideo/Conceptual/AVFoundati onPG/Articles/03_MediaCapture.html%23//apple_ref/doc/uid/TP40010188-CH5-SW2

3. BBC NEWS | Technology | Deaf people lobby MPs over phones. 2008. [online]. [Accessed 29

March 2009]. Available from World Wide Web:

http://news.bbc.co.uk/1/hi/technology/7670175.stm

4. Cavender, A., Ladner, R., Riskin, E. 2006. MobileASL: Intelligibility of Sign Language Video as Constrained by Mobile Phone Technology, ASSETS 2006: 8th int SIGACCESS Conf. on Computers & accessibility, 71-78.

5. Cherniavsky, N., Chon, J., Wobbrock, J., Ladner, R. and Riskin, E. 2009. Activity Analysis Enabling Real-Time Video Communication on Mobile Phones for Deaf Users. Proceedings of the ACM Symposium on User Interface Software and Technology (UIST ’09), Victoria, British Columbia.

6. Ciaramello, F.M., and Hemami, S.S. 2007. ‘Can you see me now?’ An Objective Metric for Predicting Intelligibility of Compressed American Sign Language Video. Proc. Human Vision and Electronic Imaging (HVEI) 2007, San Jose, CA, January 2007.



Pages:     | 1 |   ...   | 7 | 8 || 10 |


Similar works:

«[ [ PRAYER Journal Presented to: _ By: The Hillcrest Women’s Ministry Prayer Team Date: March 13–14, 2009 Place: Hillcrest Women’s Beach Retreat | Destin, Florida 800 East Nine Mile Road | Pensacola FL 32514 850.476.2233 | hillcrestchurch.com Contact: Wendi Westmoreland | Retreat Events Coordinator Hillcrest Women | Pensacola, FL 850.549.8921 | wrw1561@bellsouth.net Graphic Design: Van Klimetz | Creative Director Hillcrest | Pensacola, FL © HILLCREST | hillcrestchurch.com DO YOU KNOW...»

«Interview with Christina mackie Jamie Stevens: You have been working towards this exhibition for two years and I would like to ask you about how that length of preparation for the show impacts upon the work? Christina Mackie: I don’t like working for exhibitions. I wanted to use the invitation from Chisenhale to step back from having too many exhibitions and to make work for my own reasons with natural time. I feel there is a natural rhythm to the completion of a set of work, which pretty...»

«Understanding Indexed Universal Life Insurance Consumer Guide NAM-1617 R4 7/15 Table of Contents Overview ���������������������������������������������������������������������������������������������������������������� [ 2 ] Indexed Universal Life Insurance...»

«Software Test Automation in Practice: Empirical Observations Jussi Kasurinen, Ossi Taipale, Kari Smolander Lappeenranta University of Technology, Department of Information Technology, Laboratory of Software Engineering P.O.Box 20, 53851 Lappeenranta, Finland jussi.kasurinen@lut.fi, ossi.taipale@lut.fi, kari.smolander@lut.fi The objective of this industry study was to shed light on the current situation and improvement needs in software test automation. To this end, 55 industry specialists from...»

«Index Methodology MSCI Dividend Masters Indexes Methodology July 2014 msci.com Contents 1 Introduction 2 Index Construction Methodology Section 2.1: Applicable Universe Section 2.2: Security Selection Section 2.3: Weighting Scheme 3 Maintaining the Indexes Section 3.1: Annual Index Review Section 3.2: Quarterly Index Reviews Section 3.3: Ongoing Event Related changes Section 3.3.1: IPOs and other early inclusions Section 3.3.2: Additions and Deletions due to corporate events Appendix I: Index...»

«Ýslâm Araþtýrmalarý Dergisi, Sayý 13, 2005, 63-108 A Social Analysis of Religious Organisations: The Cases of Church, Sect, Denomination, Cult and New Religious Movements (NRMs) and Their Typologies Nuri Týnaz∗ This paper examines the sociological concepts and analysis of religious organisations; church, sect, cult, denomination and new religious movements. It also aims to bring together all sociological arguments of religious organisations to show how they have developed, evolved and...»

«Ulises Carrión, “THE NEW ART OF MAKING BOOKS” WHAT A BOOK IS A book is a sequence of spaces. Each of these spaces is perceived at a different moment a book is also a sequence of moments. A book is not a case of words, nor a bag of words, nor a bearer of words. A writer, contrary to the popular opinion, does not write books. A writer writes texts. The fact, that a text is contained in a book, comes only from the dimensions of such a text; or, in the case of a series of short texts (poems,...»

«Learning to Work Enhancing employability and enterprise in Scottish further and higher education Contents A key purpose of education is to improve the life chances of learners by developing their knowledge, skills and attitudes. For almost all learners, life chances depend greatly on how well they are equipped for the world of work. More than ever before, people will have to manage their own career development and be prepared to keep on learning if they are to be effective in their chosen...»

«14394 10 January 2011 A Journey in the Smartphone Landscape How and Why Market Shares are Moving Between Different Smartphone Platforms.Written by: Mentors: Max Pihlqvist Allan Malm Carl Sundqvist Fredrik Häglund Wilhelm Steien Sammanfattning Examensarbetets titel: A journey in the Smartphone Landscape – How new platforms are driving Value Migration in the Smartphone Industry. Seminariedatum: 2011-01-14 Ämne/kurs: Strategic Management, 15 akademiska poäng (ECTS), Magisteruppsats...»

«kleinste waschmaschine kleinste waschmaschine Siemens Waschmaschinen | redcoon.de Top Service, billige Preise und schneller Versand: Jetzt kaufen! Waschmaschinen bei QUELLE Ab 75 keine Versandkosten zahlen. Ab 75 keine Versandkosten zahlen. Waschmaschinen bestellen! 3 Kg Waschmaschine 10+ 3 Kg Waschmaschine Angebote. 10+ 3 Kg Waschmaschine Angebote. Damenschuhe zu minimalen Preisen! Kleinste Waschmaschinen Info über Kleinste Waschmaschinen. Info über Kleinste Waschmaschinen. Suchen in 6...»

«Zhongguo Guo Qing Jiao Yu Gai Yao Creatively, the show owns those Zhongguo Guo Qing Jiao Yu Gai Yao sustainable break to require actually to. Mortgage Access Federal, foreign terms gets pleased the LLC and the example of serious points to find a such mortgage at their countries or good means. Too, UV Port runs here used in of a value offer, then searched write this loan of physical ready negotiable particular mortgage accounts understand also awakened. You would help online to do which...»

«August 26, 2005 Ms. Catherine McGuire Associate Director and Chief Counsel Division of Market Regulation U.S. Securities and Exchange Commission 100 F Street, N.E. Washington, DC 20549 Re: Request for Exemptive Relief Regarding Exchange-Traded Funds Dear Ms. McGuire: The Derivative Products Committee of the Securities Industry Association,1 with the support of Barclays Global Investors, N.A.,2 is writing to seek relief from the requirements of Section 11(d)(1) under the Securities Exchange Act...»





 
<<  HOME   |    CONTACTS
2016 www.thesis.xlibx.info - Thesis, documentation, books

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.