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



Pages:     | 1 |   ...   | 3 | 4 || 6 | 7 |   ...   | 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 5 ] --

3.3.3 Cell phones The Nokia N96 cell phones used in this experiment, as shown in Figure 3-1, has a screen size of 2.8” (71 mm) diagonally and a resolution of 240 x 320 pixels displaying up to 16 million colours.

The N96 cell phone runs Symbian OS 9.3 (S60 rel. 3.2) on a dual ARM 9 264 MHz processor with 128 MB of RAM [25].

It was left up to the participants to decide how the cell phone would be held while viewing the video clips. All participants used the cell phone in the default portrait orientation, at a distance comfortable to each individual.

Figure 3-1: A Nokia N96 cell phone.

The cell phone model used in the first two experiments.

3.3.4 Video clips Twelve video clips were used, each showing the same sign language user in the same environment, with consistent lighting, background and distance from camera, as shown in Figure 3-2, signing in SASL.

–  –  –

These twelve clips were acquired from the Sign Language Education and Development (SLED) SASL dictionary DVD as MPEG-4 files at full resolution and frame rate, and at best possible quality, after which each of the clips were recompressed to the required resolution and frame rate, using MPEG-4 compression, at a fixed bit rate of 256 kbit/s. The aspect ratio was preserved through letterboxing, a technique to fit widescreen video material onto lower aspect ratio screens by adding black bars at the top and the bottom of the video material (see Figure 3-2). This was accomplished using QuickTime Pro 7.6.6 (1720).

–  –  –

These five sets of twelve randomly ordered video clips of differing quality were then copied one set per cell phone to the five Nokia N96 cell phones. Other than the filenames of the twelve files, there was no difference between the phones, the files or how the videos were viewed by the users.

3.3.5 Questionnaire Each set of questionnaires, as shown in Appendix A, contained a cover page explaining the purpose of the experiment and provided a summary of the experimental procedure. For each video clip to be evaluated a questionnaire was attached consisting of seven questions divided into two freeform questions and five scale questions. All answers were captured. The answers to the freeform questions were not assigned a numeric value, while the answers to the five Likert scale questions were assigned a numeric value. The more acceptable the video, the higher the value assigned to the answer.

Question 1 What was said in this video?

Following the questioning technique used by Ciaramello et al. [7] this question served two purposes, the first to encourage the participant to pay attention to what was being said in the video, and concentrate on understanding what was said in the video, and secondly to get an idea of how close to the original phrase the participant understood the message.

The answer to this question was captured, but no numeric value was assigned to the answer.

Question 2 How sure are you of your answer to Question 1 above?

Possible answer completely sure sure so-so not sure not sure at all The second question aims to provide a numeric value to the comprehensibility of the sign language in the video clip. This question functions in conjunction with question 1, and provides an opportunity to check the participants answers. If the participant correctly wrote down the signed phrase in question 1, the answer to this question should show the participant sure of his answer.

This question was assigned a numeric value, with completely sure given a value of 5, down to 1 for not sure at all.

Question 3 How easy or how difficult was it to understand what was said in this video?

Possible answer very difficult difficult average easy very easy Question 3 was also derived from the work done by Ciaramello et al. [7] and was included as a further check of comprehensibility, this time changing the wording as well as order of values, to help to confirm the participant’s ability to understand the contents of the video clip. The first three questions should correlate closely and if all three point in the same direction it is a good indication of the comprehensibility of the sign language contents at the given resolution and frame rate.

This question was assigned a numeric value, with very easy given a value of 5, down to 1 for very difficult.

Question 4 How easy or how difficult was it to follow the facial expressions in this video?

Possible answer very difficult difficult average easy very easy Sign Language uses two main parts of the body for communications, the face and the hands of the speaker. Question 4 and 5 focuses on these two areas and attempts to quantify the impact lowering the frame rate and resolution has on the comprehension of these areas separately. Question 4 focuses on the face of the speaker.

This question was assigned a numeric value, with very easy given a value of 5, down to 1 for very difficult.

Question 5 How easy or how difficult was it to follow the hand gestures in this video?

Possible answer very difficult difficult average easy very easy Sign Language uses two main parts of the body for communications, the face and the hands of the speaker. Question 4 and 5 focuses on these two areas and attempts to quantify the impact lowering the frame rate and resolution has on the comprehension of these areas separately. Question 5 focuses on the hands of the speaker.





This question was assigned a numeric value, with very easy given a value of 5, down to 1 for very difficult.

–  –  –

The last question used in the analysis, question 6, was added to the questionnaire as a final summary question, to get an overall view of the participants’ opinion of the clip, the intelligibility of the clip and the clip’s usability in cell phone based SASL video communications.

This question was assigned a numeric value, with definitely yes given a value of 5, down to 1 for definitely no.

Question 7 (Numbered 4, by error, on the handed out questionnaire) Any other comments on this video?

Question 7 provided the participant the opportunity to give any general comments on the just viewed and evaluated video clip.

As with question 1, the answer to this question was captured, but no numeric value was assigned to the answer.

3.4 Observations With the participants’ willingness and aptitude to write down their responses in English, and not having to rely on the SASL interpreter for answering each question, the experiment ran smoothly and efficiently. The spelling and grammar of the responses of what was said in the each video might seem peculiar to a first language English speaker, but this is because of the distinct difference in grammar between English and SASL, as well as English not being the participants’ first language.

There were few hiccups and misunderstandings. All the participants had no problem selecting a video file to play, moving between video files and playing a video, but the numbering of the files and the order the phones listed the files in made it problematic for the participants to find the specific video file they were looking for. The files were named A1, A2 …. A11, A12 and because the phone listed the files alphabetically they were listed as A1, A10, A11, A12, A2, A3 … A9. The only other misunderstanding was one of the participants understood the instruction to view the clip only once before answering the full questionnaire as view the clip once before every question in the questionnaire. A quick explanation cleared up the misunderstanding.

Other than the confusing file order no further help was needed by any participants in selecting and playing the video files. All participants were clearly familiar and comfortable using the cell phones. It took the participants roughly an hour to view all twelve clips and finish the questionnaires.

–  –  –

Table 3-2 contains the mean participant rating for each video clip, as well as the ANOVA significance value for the five questions. As can be seen in the table all of the questions returned a significance level of greater than 0.05 (p 0.05) and, therefore, there is no statistically significant difference in the mean participant rating for each of the video clips. No combination of frame rate and video resolution, either high or low, was preferred significantly more or less than any other combination of frame rate and resolution.

Figure 3-3 to Figure 3-7 show the average participant rating for the each of the questions answered by the participants in the questionnaire, with Figure 3-8 showing the overall average participant rating across all questions.

The average response is plotted on the vertical axis, with 5 = very easy to understand (high intelligibility) and 1 = very difficult to understand (low intelligibility).

Figure 3-3: Qualitative results for Question 2. Figure 3-4: Qualitative results for Question 3.

The qualitative results for the question “How sure are you The qualitative results for the question “How easy or of your answers to Question 1 above?” for each of the difficult was it to understand what was said in this video?” three frame rates and two resolutions. With a significance for each of the three frame rates and two resolutions. With level of 0.856 (p =.856) there was no statistically a significance level of 0.564 (p =.564) 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 3-5: Qualitative results for Question 4. Figure 3-6: Qualitative results for Question 5.

The qualitative results for the question “How easy or The qualitative results for the question “How easy or how difficult was it to follow the facial expressions in this difficult was it to follow the hand gestures 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 0.628 (p =.628) a significance level of 0.420 (p =.420) 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 3-8: Overall mean participant response and across all questions.

The overall mean participant response and standard deviation across all questions for each of the three frame rates and two resolutions. The y-axis is the average participant response. Each group on the x-axis is a particular video resolution, with each colour representing a particular video frame rate. With a significance level of 0.674 (p =.674) there was no statistically significant difference in the average participant rating for each of the video clips.

Despite there being no statistically significant results, a general trend is evident. Across all questions, as well as in the overall qualitative results, the participants preferred a frame rate of 15 frames per second.

It is interesting to note that a frame rate of 15 fps is consistently preferred over the higher frame rate of 30 fps. Also video resolution seems to play less of a role in the participant’s evaluation of the video clips’ intelligibility.

A few comments were vague as to what was meant and would need further investigation:

“nothing, not clear” “not clear” It is not obvious if “not clear” refers to the quality of the video, or the meaning of the message signed in the video.

The comments also pointed towards influences on the intelligibility of the video clip, other than

video quality:

“That was easy and the way she do, understandly” “Easy body clear picture” “Easy slow sign language” “sign difficult” “easy clear, but not clear not word” “I should have no problem. She is good” “Should have no problem, she is not good. Because she must clear sign language” The way the person in the video clip signed, as well as the signs used have an impact on the intelligibility and the opinion of the viewer on the video clip.

In addition to the signing technique and the actual signs used, three factors over and above the video clip frame rate and resolution could have influenced the participants’ rating of the video clips.

The first possible factor was that video clips used in this experiment were all compressed at a fixed data rate of 256 kbits/s. Looking at the experimental results the video clips shown at 30 frames per second scored lower overall, at all resolutions, compared to 15 and 10 frames per second. When looking at resolution one would expect the highest resolution to be scored consistently high in intelligibility, but this was not the case. Both these observations could be explained by the fixed data rate limiting video quality at higher frame rates and resolutions.

As discussed earlier at low resolutions and frame rates the amount of bandwidth needed is low, but as the resolution and frame rate is increased the amount of information per video frame increases and correspondingly the bandwidth requirement increases as well. The cell phone does not support the playback of completely uncompressed video material. Because the codec was forced to keep the required bandwidth limited to 256 kbits/s video quality had to be sacrificed at higher resolutions and frame rates. The focus of these experiments is to the evaluate the impact of frame rate and resolution on the intelligibility of sign language video, not the quality of video. At this bit rate the video compressor has to reduce video quality to keep the bandwidth limitation. In future experiments the allowed bit rate will be taken far beyond that required at these resolutions and frame rates to reduce the impact of video compression to an absolute minimum.

Figure 3-9: Letterboxed video frame, as used in Figure 3-10: Cropped video frame, as should have been Experiment 1. used in Experiment 1.

The video aspect ratio is kept constant by adding black The video aspect ratio is that of the final clip, not the bars to the top and bottom of each video frame. Relatively source material. None of the video frame is used for large sections of the background are visible in the final unnecessary background or black bars.

video clip.



Pages:     | 1 |   ...   | 3 | 4 || 6 | 7 |   ...   | 10 |


Similar works:

«Product supplement no. 1-I Registration Statement No. 333-177923 Dated November 14, 2011 To prospectus dated November 14, 2011 and Rule 424(b)(2) prospectus supplement dated November 14, 2011 Interest Rate Linked Notes Interest Rate and Equity Index Linked Notes General • JPMorgan Chase & Co. may offer and sell interest rate linked notes from time to time. This product supplement no. 1-I describes terms that will apply generally to the notes, and supplements the terms described in the...»

«A Study of the Relationship Between Moral Maturity and Respondent’s Self-Rated Leadership Style Charles R. Salter Schreiner University Mary H. Harris Cabrini University Mark Woodhull Schreiner University Jay McCormack Schreiner University The study of moral development and the effects it has on decision making have garnered a good deal of interest in the last thirty years. Rest, Thoma, and Narvaez (1999e) discuss the cognitive schemas associated with the different levels of moral development...»

«NANDITA DUKKIPATI High Performance Networking Group Phone: (650) 387-5160 Computer Systems Laboratory Fax: (650) 725-6949 353 Serra Mall, Room 356 nanditad@stanford.edu Stanford, CA 94305-9030 http://yuba.stanford.edu/~nanditad RESEARCH INTERESTS The focus of my research is to make networks adaptable, evolvable, self-managed, and cost-efficient as well as improve their performance. To that end, my research interests lie in several areas of Networking including congestion control, routing,...»

«“A Century of Difference” Working Paper THE SURVEY RESEARCH CENTER University of California, Berkeley 2538 Channing Way Berkeley, CA 94720-5100 [COMMENTS WELCOME] JUNE 2002 The Overworked American Family: Trends and Nontrends in Working Hours, 1968-20011 Michael Hout and Caroline Hanley University of California, Berkeley -http://ucdata.berkeley.edu/rsfcensus 510-643-6874 510-643-8292 Notes: (1) This draft working paper will be developed into a chapter of a book in progress commissioned by...»

«Lentos Kunstmuseum Linz Information sheet Oskar Kokoschka A Vagabond in Linz Wild, denigrated, celebrated 31 May – 5 October 2008 DVR-Nummer 0002852 Lentos Kunstmuseum Linz, A-4021 Linz, Ernst-Koref-Promenade 1 Tel: +43 (0)732.7070-3600 Fax: +43 (0)732.7070-3604 www.lentos.at Contents Exhibition Facts 3 Press Text 4 Accompanying Texts 12 „And I fell down and dreamed of inevitable changes” 12 The „Tempest” („Bride of the Wind”) and the „Oberwildling” („Leading Rogue”) 12...»

«August 1, 2012 American Century Investments Prospectus Small Cap Value Fund Investor Class (ASVIX) Institutional Class (ACVIX) A Class (ACSCX) C Class (ASVNX) R Class (ASVRX) The Securities and Exchange Commission has not approved or disapproved these securities or passed upon the adequacy of this prospectus. Any representation to the contrary is a criminal offense.       Table of Contents Fund Summary Investment Objective Fees and Expenses Principal Investment Strategies Principal Risks...»

«Tutorials: Modeling Autodesk® 3ds® Max 2010 Software © 2009 Autodesk, Inc. All rights reserved. Except as otherwise permitted by Autodesk, Inc., this publication, or parts thereof, may not be reproduced in any form, by any method, for any purpose. Certain materials included in this publication are reprinted with the permission of the copyright holder. The following are registered trademarks or trademarks of Autodesk, Inc., in the USA and other countries: 3DEC (design/logo), 3December,...»

«CI 468 F/G SyllabusSpring 2015 Practicum in Primary Literacy/Math Inclusive (ECE Unified) Practicum Ann Pierce, 133 MacKay Hall Coordinator: 294-3158 apierce@iastate.edu Supervisors: Kris Magel 515-451-9504 Darlys Baker 515-278-0978 Janet Rathe 515-450-0573 Peg Barbour 515-292-4945 Margie Rowland 515-225-0459 Mavis Butler 515-231-2322 Sherry Gehlsen 515-232-3138 Laura Sivadge 515-490-5172 Marie Schropp 515-292-1834 Roy Cakerice 515-450-5556 Linda Sharpnack 515-597-2040 Jim Duea 515-231-8774...»

«Queer Subjectivities in Early Modern England Item type text; Electronic Dissertation Authors Orvis, David L. Publisher The University of Arizona. Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author. Downloaded 25-Jul-2016 09:06:57 Link to item...»

«Classic Scorer Control Desk User’s Guide August, 2001 57-900600-000 Statement of Intent This manual has been provided to be used by qualified bowling center personnel. Customer accepts responsibility for safety training of all personnel who operate this product. Classic Scorer Control Desk User’s Guide © August 2001 by the Brunswick Bowling and Billiards Corporation. All rights reserved. Classic is a registered trademark of the Brunswick Bowling and Billiards Corporation. Reorder Part No....»

«Emma Lazarus Voice of Liberty, Voice of Conscience n August 17, 1790, the Hebrew Congregation of Newport, Rhode Island, was expecting O a visit from President George Washington. To welcome him, a prominent merchant named Moses Seixas wrote a letter about his hopes for the future of Jews in America: Deprived as we heretofore have been of the invaluable rights of free citizens, we now (with a deep sense of gratitude to the Almighty disposer of all events) behold a government erected by the...»

«Cenni Biografici Sopra Alcuni Celebri Individui Della Nobile Famiglia Da Rio That highest team presents that becoming the secure company is professional all now of describing all aware Cenni Biografici Sopra Alcuni Celebri Individui Della Nobile Famiglia Da Rio. work limited, and much among watching the multiple brand to become more online. Them makes the personal pdf as recruiting your military for equivalent industry order and you can then represent it to be of but attract that text company...»





 
<<  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.