Abstract DARWIN (http://darwin.eckerd.edu) is a computer program developed to facilitate the identification of individual dolphins from photographs of.

Slides:



Advertisements
Similar presentations
1 CASUS Authoring System 07/2010 E-Learning & E-Teaching Welcome to the CASUS Authoring System!
Advertisements

Support.ebsco.com EBSCOhost Collection Manager Selector Account Functions Tutorial.
Publishers Web Sites Standard Features. Objectives Access publishers websites Identify general features available on most publishers websites Know how.
OptiShip ® Multi-carrier Shipping System. OptiShip ® customers save on average 13.6% of parcel shipping costs… OptiShip ® is a comprehensive system that.
Tutorial EBSCOadmin User Groups support.ebsco.com.
STUDENT GUIDE. Go to the PUC Homepage located at From the Student drop-down menu, move cursor over the myPUC link and click myPUC Portal.
IMAGE Semi-automatic 3D building extraction in dense urban areas using digital surface models Dr. Philippe Simard President SimActive Inc.
Premier Director Document Imaging
Feature Detection and Outline Registration in Dorsal Fin Images A. S. Russell, K. R. Debure, Eckerd College, St. Petersburg, FL Most Prominent Notch analyze.
6 th Annual Focus Users’ Conference 6 th Annual Focus Users’ Conference Scheduling Requests and Request Reports Presented by: Sara Sayasane Presented by:
Introduction Researchers involved in the study of dolphin biology, such as migrations, ranging patterns and social association patterns, frequently use.
Microsoft Word 2003 Tutorial 2 – Editing and Formatting a Document.
Case Tools Trisha Cummings. Our Definition of CASE  CASE is the use of computer-based support in the software development process.  A CASE tool is a.
References [1] DARWIN. Eckerd College. darwin.eckerd.edu [2] FinScan. Texas A&M University. [3J M. S. Prewitt and M.L. Mendelsohn, The analysis of cell.
6 th Annual Focus Users’ Conference 6 th Annual Focus Users’ Conference Accounts Receivable Presented by: Robert Myers Presented by: Robert Myers.
ASENT_MTTR.PPT Maintainability Predictions Last revised 8/11/2005.
Prototyping. Horizontal Prototyping Description of Horizontal Prototyping A Horizontal, or User Interface, Prototype is a model of the outer shell of.
Automating Tasks With Macros
Creating Custom Forms. 2 Design and create a custom form You can create a custom form by modifying an existing form or creating a new form. Either way,
Automating Tasks With Macros. 2 Design a switchboard and dialog box for a graphical user interface Database developers interact directly with Access.
1 Computing for Todays Lecture 4 Yumei Huo Fall 2006.
Cambodia-India Entrepreneurship Development Centre - : :.... :-:-
Access Tutorial 3 Maintaining and Querying a Database
Hands-On Microsoft Windows Server 2008 Chapter 11 Server and Network Monitoring.
Adobe Forms THE FORM ELEMENT PANEL. Creating a form using the Adobe FormsCentral is a quick and easy way to distribute a variety of forms including surveys.
Created by: Ian Osborn. Possibilities Of Movie Maker Windows Movie Maker allows users to organize and add effects to media clips that ordinarily would.
ARCHIBUS Log On Instructions. Log Into ARCHIBUS Web Central Log In Screen 1.Open your Internet browser. 2.Enter the URL to view the ARCHIBUS Login Page.
XP New Perspectives on Microsoft Access 2002 Tutorial 41 Microsoft Access 2002 Tutorial 4 – Creating Forms and Reports.
39 Copyright © 2007, Oracle. All rights reserved. Module 39: Siebel Task UI Siebel 8.0 Essentials.
U.S. D EPARTMENT OF H OUSING AND U RBAN D EVELOPMENT U.S. D EPARTMENT OF H OUSING AND U RBAN D EVELOPMENT LOCCS M ANUAL L ARRY W OOD A PRIL 9, 2014.
XP New Perspectives on Microsoft Access 2002 Tutorial 51 Microsoft Access 2002 Tutorial 5 – Enhancing a Table’s Design, and Creating Advanced Queries and.
Where Innovation Is Tradition SYST699 – Spec Innovations Innoslate™ System Engineering Management Software Tool Test & Analysis.
Support.ebsco.com EBSCOhost Basic Searching for Academic Libraries Tutorial.
1 4HPlus – Retrieving Information March Retrieving Information The real value of any information based software is in the data and reports that.
0 eCPIC Admin Training: Custom Calculated Fields These training materials are owned by the Federal Government. They can be used or modified only by FESCOM.
Creating your own form from scratch.. To create a custom form, you can modify an existing form or design and create a form from scratch. In either case,
Clearly Visual Basic: Programming with Visual Basic 2008 Chapter 4 I Need a Tour Guide.
Engr. M. Fahad Khan Lecturer Software Engineering Department University Of Engineering & Technology Taxila.
DARWIN - Dolphin Photo-identification Software Adaptations to Digital Camera Acquisition and Increased Matching Accuracy K. R. Debure, J. H. Stewman, S.
Introduction of Geoprocessing Topic 7a 4/10/2007.
2 Session 22 Tools For Schools – FSA Assessments and the ISIR Analysis Tool Software.
COMP106 Assignment 2 Proposal 1. Interface Tasks My new interface design for the University library catalogue will incorporate all of the existing features,
SRM Free Text Carts SRM_SHO_303 SRM Free Text Carts.
Enhancing - Vuser Scripts In HP LoadRunner >>>>>>>>>>>>>>>>>>>>>>
CREATING TEMPLATES CREATING CUSTOM CHARACTERS IMPORTING BATCH DATA SAVING DATA & TEMPLATES CREATING SERIES DATA PRINTING THE DATA.
ISV Innovation Presented by ISV Innovation Presented by Business Intelligence Fundamentals: Data Cleansing Ola Ekdahl IT Mentors 9/12/08.
0 eCPIC User Training: Dependency Mapper These training materials are owned by the Federal Government. They can be used or modified only by FESCOM member.
Page 1 Non-Payroll Cost Transfer Enhancements Last update January 24, 2008 What are the some of the new enhancements of the Non-Payroll Cost Transfer?
Basic Editing Lesson 2.
GISMO/GEBndPlan Overview Geographic Information System Mapping Object.
MetaLib 4 User Guide. 2 MetaLib 4 Access MetaLib at: – MetaLib may be used at two different levels –
Configuration Management and Change Control Change is inevitable! So it has to be planned for and managed.
XP New Perspectives on Microsoft Office Access 2003 Tutorial 10 1 Microsoft Office Access 2003 Tutorial 10 – Automating Tasks With Macros.
User Interface Components Lecture # 5 From: interface-elements.html.
An Introduction to Forms. The Major Steps of a MicroSoft Access Database  Tables  Queries  Forms  Macros  Reports  Modules On our road map, we are.
Word Create a basic TOC. Course contents Overview: table of contents basics Lesson 1: About tables of contents Lesson 2: Format your table of contents.
Introduction of Geoprocessing Lecture 9 3/24/2008.
XP New Perspectives on Microsoft Office Access 2003, Second Edition- Tutorial 6 1 Microsoft Office Access 2003 Tutorial 6 – Creating Custom Forms.
Basic Navigation in Oracle R12 BY: Muhammad Irfan.
6 th Annual Focus Users’ Conference 6 th Annual Focus Users’ Conference ERP Fixed Assets ERP Fixed Assets Presented by: Will Brock ERP Programmer PowerPoint.
Processing Manual Invoices July, 2015 The new XR-7.
Text2PTO: Modernizing Patent Application Filing A Proposal for Submitting Text Applications to the USPTO.
HTBN Batches These slides are intended as a starting point for further discussion of how eTime might be extended to allow easier processing of HTBN data.
MS WORD INFORMATION TECHNOLOGY MANAGEMENT SERVICE Training & Research Division.
The Next Step Hudson Fare Files 102 – Import & upload Rev. 10/14.
Emdeon Office Batch Management Services This document provides detailed information on Batch Import Services and other Batch features.
Human Computer Interaction Lecture 21 User Support
Presentation transcript:

Abstract DARWIN ( is a computer program developed to facilitate the identification of individual dolphins from photographs of their dorsal fins. Users query a digital catalog of photographs of individual dolphins to produce a rank ordered display of images identified as most similar to the query image. Thus, the software attempts to reduce the number of images a researcher must examine to identify a particular individual. Software development without early and continuing input from end users can lead to a final product which is cumbersome and non intuitive, if useful at all. Complex software such as DARWIN must be tested in a manner that emulates its actual usage as completely as developers can manage. Usage environments can differ significantly with respect to database size, image size and quality, and perhaps most significantly established protocols for manually processing data and user preferences related to computer usage. Something a simple as an awkward key or button sequence can make the accomplishment of a desired outcome pleasant or tedious. Moreover, as technology changes and research groups adopt different standards for gathering and maintaining field data, the software must also adapt. Always, the goal is to facilitate the efficient processing of photographs, to the extent possible. This discussion of our experience with the design and development of DARWIN focuses on feedback from users of initial beta releases of the software and from our collaboration with the Eckerd College Dolphin Project. We detail some of the significant changes and features suggested by current and potential users. Initially, DARWIN was developed with the intent of reducing the tedium associated with the manual identification of individual dolphins using a print catalog. Most groups were using slide film to photograph dolphins in the wild and the bottleneck of scanning the slides was a deterrent to use of an automated method of identification. As the price of digital cameras has dropped and the available resolution of those cameras has improved to that of film quality, the possibility of using computer software tools to aid in the identification of individuals has become much more appealing and widespread. Outline generation Most dolphin research groups currently use off the shelf software to display, organize, enhance and/or transform their digital images for comparison. In order to offer an attractive alternative to these products, DARWIN must reduce the time it takes to identify an individual (on average) and must not require significant additional work from its users; otherwise the useful "tool" becomes an onerous "task". The DARWIN software uses dorsal fin outlines in the comparison of one fin to another. The detection of a fin outline in an image can be initiated by using the mouse to indicate its general image placement. This placement is then refined automatically with the use of active contours. Even though this semi-automated process works well and requires only a few seconds of user input, the initial task of tracing outlines can be laborious over a large collection of images. Therefore, all beta releases of the DARWIN software have included a fully automatic method of generating fin outlines. If this autotrace fails, the user can always retrace using the semi-automated approach. In our testing of 302 images, approximately 68% of the autotraced outlines were useful. Feedback from one group who experimented with the software had much less success. Our examination of their images, showed that simply cropping to minimize foreground glare significantly improved results. We conclude that clearer documentation about how to make the software work well will be invaluable. Acknowledgements The authors would like to thank Dr. Shannon Gowans and the students of the ECDP, the folks at the Harbor Branch Oceanographic Institute and the Duke Marine Lab, and Dr. Hal Whitehead for their considerable feedback and suggestions. We would also like to thank the National Science Foundation for funding of this research under grant number DBI Accuracy As the software has continued in testing, we have made several minor adjustments to the fin registration (mapping) and ranking code in attempts to further improve the quality of the mappings and the accuracy of the final rankings of the match. The tip position is now less constrained in the iterative mapping process and this has led to results as follows. A test set of fifty “unknown” individuals was compared against a catalog of 200 individuals. Forty one of the fifty queries produced rankings in the top 5% of the catalog. Of these, 21 queries ranked the correct fin first. On average, the query placed the correct fin in the top 5% of the catalog, with a median ranking in the top 1%. Flipping/zooming/overlay In order to enable an automatically generated outline to be created, the image may be contrast enhanced and cropped to include only the dorsal fin. Although these enhancements may improve the ability of the software to extract an outline, they may hide or de-emphasize features used by biologists for comparison of the dorsal fins. Thus, when displaying the match results, DARWIN displays the modified, cropped images by default, but users may alternatively display the original image for comparison. Also, the orientation of the displayed fins can be changed for ease of comparison. Efforts are currently underway to enable the sighted fin to be displayed such that it appears to have been photographed from the same angle as the catalog fin to which it is being compared. Suggestions have been made to combine this utility with the ability to overlay the two aligned images to facilitate comparison. Software design and user feedback in the development of DARWIN John H. Stewman and Kelly R. Debure, Eckerd College. Bigger pictures Viewing the results of a match (the ranked list of catalog images) is enabled in a side by side display of the unknown sighting image and each image in the catalog. Our discussions with Dr. Shannon Gowans of the ECDP made one thing clear: when displaying images of dorsal fins, bigger is better. The current software has a match results window that has been completely redesigned to maximize the screen space allocated to the side by side display of the images. As catalogs become large, the click of a single button to view the next image can become tiresome and monotonous. It has been suggested, again by Dr. Gowans, that a slide show viewing option be added to the window which displays the match results. This function is being incorporated. Figure 3: The window which displays the rank ordered list of matching images has been redesigned to prioritize display of larger images, thus facilitating their comparison. Figure 1: The window above shows an image and its outline which has been traced during one interactive session and reloaded during another. catalogs against themselves or one another. To prevent the need to trace the outline twice, once for the creation of the catalog and once for the creation of the query outline, the current software provides the ability to save a traced fin for later use. The trace of the outline and its associated image can either be loaded to create a catalog or to search for a match in the catalog. A trace can also be saved for batch processing as described below. Saving a trace In addition to identifying individuals from a sighting against a database of known individuals, dolphin researchers occasionally want to compare entire Figure 2: In the window above, a queue of three outlines has been created for batch processing of matches against the catalog. Progress bars indicate the percentage of the matching which has been completed. Running multiple matches in queue mode Once a catalog has been created and a trace has been extracted from the dorsal fin in a sighting image, the process of matching can proceed. Although the software makes individual comparisons fairly quickly, generating a ranked list of matches against a large database can take minutes rather than seconds. On a typical 1.6 GHz laptop running Windows XP, the software takes about 20 seconds to create the ranked listing from a database of 200 fins. Initial users of the beta release inquired about the ability to submit the match requests as a batch process and reload the results for viewing later. This feature has been added.