Faceted Metadata in Image Search Browsing Using Words

Faceted Metadata in Image Search & Browsing Using Words to Browse a Thousand Images Ka-Ping Yee, Kirsten Swearingen, Kevin Li, Marti Hearst Group for User Interface Research UC Berkeley CHI 2003 Research funded by: NSF CAREER Grant IIS-9984741 IBM Faculty Fellowship

Outline • How do people search and browse for images? • Current approaches: – Keywords – Spatial similarity • Our approach: – Hierarchical Faceted Metadata – Very careful UI design and testing • Usability Study • Conclusions M. Hearst Faceted Metadata in Search

How do people want to search and browse images? Ethnographic studies of people who use images intensely: – Finding specific objects is easy – Find images of the Empire State Building – Browsing is difficult – People want to use rich descriptions. M. Hearst Faceted Metadata in Search

Ethnographic Study • Markkula & Sormunen ’ 00 – Journalists and newspaper editors – Choosing photos from a digital archive • Searching for specific objects is trivial • Stressed a need for browsing • Photos need to deal with themes, places, types of objects, views – Had access to a powerful interface, but it had 40 entry forms and was generally hard to use; no one used it. M. Hearst Faceted Metadata in Search

Markkula & Sormunen ’ 00 M. Hearst Faceted Metadata in Search

Query Study • Armitage & Enser ’ 97 – Analyzed 1, 749 queries submitted to 7 image and film archives – Classified queries into a 3 x 4 facet matrix • Rio Carnivals: Geo Location x Kind of Event – Concluded that users want to search images according to combinations of topical categories. M. Hearst Faceted Metadata in Search

Ethnographic Study • Ame Elliot ’ 02 – Architects • Common activities: – Use images for inspiration • Browsing during early stages of design – Collage making, sketching, pinning up on walls • This is different than illustrating powerpoint • Maintain sketchbooks & shoeboxes of images – Young professionals have ~500, older ~5 k • No formal organization scheme – None of 10 architects interviewed about their image collections used indexes • Do not like to use computers to find images M. Hearst Faceted Metadata in Search

Current Approaches to Image Search • Keyword based – Web. Seek (Smith and Jain ’ 97) – Commercial web image search systems – Commercial image vendors (Corbis, Getty) – Museum web sites M. Hearst Faceted Metadata in Search

Current Approaches to Image Search • Using Visual “Content” – Extract color, texture, shape • QBIC (Flickner et al. ‘ 95) • Blobworld (Carson et al. ‘ 99) • Piction: images + text (Srihari et al. ’ 91 ’ 99) – Two uses: • Show a clustered similarity space • Show those images similar to a selected one – Usability studies: • Rodden et al. : a series of studies • Clusters don’t work; showing textual labels is promising. M. Hearst Faceted Metadata in Search

Rodden et al. , CHI 2001 M. Hearst Faceted Metadata in Search

How Best to Support Browsing? • To support serendipity, want to view images that are related along multiple dimensions. • But clusters are not comprehensible. • Instead, allow users to “steer” through the multi-dimensional category space in a flexible manner. M. Hearst Faceted Metadata in Search

Some Challenges • Users don’t like new search interfaces. • How to show lots more information without overwhelming or confusing? M. Hearst Faceted Metadata in Search

Our Approach • Integrate the search seamlessly into the information architecture. – Use proper HCI methodologies. • Use faceted metadata: – More flexible than canned hyperlinks – Less complex than full search – Help users see where to go next and return to what happened previously M. Hearst Faceted Metadata in Search

Metadata: data about data Facets: orthogonal categories Geo. Region M. Hearst + Time/Date + Topic Faceted Metadata in Search

Hierarchical Faceted Metadata Example: Biological Subject Headings 1. Anatomy [A] 2. Organisms [B] 3. Diseases [C] 4. Chemicals and Drugs [D] 5. Analytical, Diagnostic and Therapeutic Techniques and Equipment [E] 6. Psychiatry and Psychology [F] 7. Biological Sciences [G] 8. Physical Sciences [H] 9. Anthropology, Education, Sociology and Social Phenomena [I] 10. Technology and Food and Beverages [J] 11. Humanities [K] 12. Information Science [L] 13. Persons [M] 14. Health Care [N] 15. Geographic Locations [Z] M. Hearst Faceted Metadata in Search

Hierarchical Faced Metadata 1. Anatomy [A] Body Regions [A 01] 2. [B] Musculoskeletal System [A 02] 3. [C] Digestive System [A 03] 4. [D] Respiratory System [A 04] 5. [E] Urogenital System [A 05] 6. [F] …… 7. [G] 8. Physical Sciences [H] 9. [I] 10. [J] 11. [K] 12. [L] 13. [M] M. Hearst Faceted Metadata in Search

Hierarchical Faceted Metadata 1. Anatomy [A] Body Regions [A 01] 2. [B] Musculoskeletal System [A 02] 3. [C] Digestive System [A 03] 4. [D] Respiratory System [A 04] 5. [E] Urogenital System [A 05] 6. [F] …… 7. [G] 8. Physical Sciences [H] 9. [I] 10. [J] 11. [K] 12. [L] 13. [M] M. Hearst Abdomen [A 01. 047] Back [A 01. 176] Breast [A 01. 236] Extremities [A 01. 378] Head [A 01. 456] Neck [A 01. 598] …. Faceted Metadata in Search

Hierarchical Faceted Metadata 1. Anatomy [A] Body Regions [A 01] 2. [B] Musculoskeletal System [A 02] 3. [C] Digestive System [A 03] 4. [D] Respiratory System [A 04] 5. [E] Urogenital System [A 05] 6. [F] …… 7. [G] 8. Physical Sciences [H] Electronics 9. [I] Astronomy 10. [J] Nature 11. [K] Time 12. [L] Weights and Measures 13. [M] …. M. Hearst Abdomen [A 01. 047] Back [A 01. 176] Breast [A 01. 236] Extremities [A 01. 378] Head [A 01. 456] Neck [A 01. 598] …. Faceted Metadata in Search

Hierarchical Faceted Metadata 1. Anatomy [A] Body Regions [A 01] Abdomen [A 01. 047] 2. [B] Musculoskeletal System [A 02] Back [A 01. 176] 3. [C] Digestive System [A 03] Breast [A 01. 236] 4. [D] Respiratory System [A 04] Extremities [A 01. 378] 5. [E] Urogenital System [A 05] Head [A 01. 456] 6. [F] …… Neck [A 01. 598] 7. [G] …. 8. Physical Sciences [H] Electronics Amplifiers 9. [I] Astronomy Electronics, Medical 10. [J] Nature Transducers 11. [K] Time 12. [L] Weights and Measures 13. [M] …. M. Hearst Faceted Metadata in Search

Hierarchical Faceted Metadata 1. Anatomy [A] Body Regions [A 01] Abdomen [A 01. 047] 2. [B] Musculoskeletal System [A 02] Back [A 01. 176] 3. [C] Digestive System [A 03] Breast [A 01. 236] 4. [D] Respiratory System [A 04] Extremities [A 01. 378] 5. [E] Urogenital System [A 05] Head [A 01. 456] 6. [F] …… Neck [A 01. 598] 7. [G] …. 8. Physical Sciences [H] Electronics Amplifiers 9. [I] Astronomy Electronics, Medical 10. [J] Nature Transducers 11. [K] Time 12. [L] Weights and Measures Calibration 13. [M] …. Metric System Reference Standard M. Hearst Faceted Metadata in Search

The Interface Design • Chess metaphor – Opening – Middle game – End game M. Hearst Faceted Metadata in Search

M. Hearst Faceted Metadata in Search

The Interface Design • Tightly Integrated Search • Supports Expand as well as Refine • Dynamically Generated Pages – Paths can be taken in any order • Consistent Color Coding • Consistent Backup and Bookmarking • Standard HTML M. Hearst Faceted Metadata in Search

What is Tricky About This? • It is easy to do it poorly – Yahoo directory structure • It is hard to be not overwhelming – Most users prefer simplicity unless complexity really makes a difference • It is hard to “make it flow” – Can it feel like “browsing the shelves”? M. Hearst Faceted Metadata in Search

Project History • Identify Target Population – Architects, city planners • Needs assessment. – Interviewed architects and conducted contextual inquiries. • Lo-fi prototyping. – Showed paper prototype to 3 professional architects. • Design / Study Round 1. – Simple interactive version. Users liked metadata idea. • Design / Study Round 2: – Developed 4 different detailed versions; evaluated with 11 architects; results somewhat positive but many problems identified. Matrix emerged as a good idea. • Metadata revision. – Compressed and simplified the metadata hierarchies M. Hearst Faceted Metadata in Search

Project History • Design / Study Round 3. – New version based on results of Round 2 – Highly positive user response • Identified new user population/collection – Students and scholars of art history – Fine arts images • Study Round 4 – Compare the metadata system to a strong, representative baseline M. Hearst Faceted Metadata in Search

New Usability Study • Participants & Collection – 32 Art History Students – ~35, 000 images from SF Fine Arts Museum • Study Design – Within-subjects • Each participant sees both interfaces • Balanced in terms of order and tasks – Participants assess each interface after use – Afterwards they compare them directly • Data recorded in behavior logs, server logs, papersurveys; one or two experienced testers at each trial. • Used 9 point Likert scales. • Session took about 1. 5 hours; pay was $15/hour M. Hearst Faceted Metadata in Search

The Baseline System • Floogle • Take the best of the existing keywordbased image search systems M. Hearst Faceted Metadata in Search

Comparison of Common Image Search Systems M. Hearst System Collection # Results Catego # /page ries? Familiar Google Web 20 No 27 Alta. Vista Web 15 No 8 Corbis Photos 9 -36 No 8 Getty Photos, Art 12 -90 Yes 6 MS Office Photos, Clip art 6 -100 Yes N/A Thinker Fine arts images 10 Yes 4 BASELINE Fine arts images 40 Yes N/A Faceted Metadata in Search

sword M. Hearst Faceted Metadata in Search

M. Hearst Faceted Metadata in Search

Evaluation Quandary • How to assess the success of browsing? – Timing is usually not a good indicator – People often spend longer when browsing is going well. • Not the case for directed search – Can look for comprehensiveness and correctness (precision and recall) … – … But subjective measures seem to be most important here. M. Hearst Faceted Metadata in Search

Hypotheses • We attempted to design tasks to test the following hypotheses: – Participants will experience greater search satisfaction, feel greater confidence in the results, produce higher recall, and encounter fewer dead ends using FC over Baseline – FC will perceived to be more useful and flexible than Baseline – Participants will feel more familiar with the contents of the collection after using FC – Participants will use FC to create multi-faceted queries M. Hearst Faceted Metadata in Search

Four Types of Tasks – Unstructured (3): Search for images of interest – Structured Task (11 -14): Gather materials for an art history essay on a given topic, e. g. • Find all woodcuts created in the US • Choose the decade with the most • Select one of the artists in this periods and show all of their woodcuts • Choose a subject depicted in these works and find another artist who treated the same subject in a different way. – Structured Task (10): compare related images • Find images by artists from 2 different countries that depict conflict between groups. – Unstructured (5): search for images of interest M. Hearst Faceted Metadata in Search

Other Points • Participants were NOT walked through the interfaces. • The wording of Task 2 reflected the metadata; not the case for Task 3 • Within tasks, queries were not different in difficulty (t’s<1. 7, p >0. 05 according to post-task questions) • Flamenco is and order of magnitude slower than Floogle on average. – In task 2 users were allowed 3 more minutes in FC than in Baseline. – Time spent in tasks 2 and 3 were significantly longer in FC (about 2 min more). M. Hearst Faceted Metadata in Search

Results • Participants felt significantly more confident they had found all relevant images using FC (Task 2: t(62)=2. 18, p<. 05; Task 3: t(62)=2. 03, p<. 05) • Participants felt significantly more satisfied with the results (Task 2: t(62)=3. 78, p<. 001; Task 3: t(62)=2. 03, p<. 05) • Recall scores: M. Hearst – Task 2 a: In Baseline 57% of participants found all relevant results, in FC 81% found all. – Task 2 b: In Baseline 21% found all relevant, in FC 77% found all. Faceted Metadata in Search

Post-Interface Assessments M. Hearst All significant at p<. 05 except simple and overwhelming Faceted Metadata in Search

Perceived Uses of Interfaces Baseline FC M. Hearst Faceted Metadata in Search

Post-Test Comparison Which Interface Preferable For: Baseline FC Find images of roses Find all works from a given period 15 16 2 30 Find pictures by 2 artists in same media 1 29 4 28 8 23 6 24 28 3 1 31 2 29 M. Hearst Faceted Metadata in Search

Post-Test Comparison Which Interface Preferable For: Baseline FC Find images of roses Find all works from a given period 15 16 2 30 Find pictures by 2 artists in same media 1 29 4 28 8 23 6 24 28 3 1 31 2 29 Overall Assessment: More useful for your tasks Easiest to use Most flexible More likely to result in dead ends Helped you learn more Overall preference M. Hearst Faceted Metadata in Search

Facet Usage • Facets driven largely by task content – Multiple facets 45% of time in structured tasks • For unstructured tasks, – – – Artists (17%) Date (15%) Location (15%) Others ranged from 5 -12% Multiple facets 19% of time • From end game, expansion from M. Hearst – Artists (39%) – Media (29%) – Shapes (19%) Faceted Metadata in Search

Qualitative Observations • Baseline: – Simplicity, similarity to Google a plus – Also noted the usefulness of the category links • FC: – Starting page “well-organized”, gave “ideas for what to search for” – Query previews were commented on explicitly by 9 participants – Commented on matrix prompting where to go next • 3 were confused about what the matrix shows – Generally liked the grouping and organizing – End game links seemed useful; 9 explicitly remarked positively on the guidance provided there. – Often get requests to use the system in future M. Hearst Faceted Metadata in Search

Study Results Summary • Strongly positive results for the faceted metadata interface. • Moderate use of multiple facets. • Strong preference over the current state of the art. – Chair of Architecture Dept: “It felt like I was browsing the shelves!” – This kind of enthusiasm is not seen in similaritybased image search interfaces. • Hypotheses are supported. M. Hearst Faceted Metadata in Search

Implementation • All open source code – Mysql database – Python web server (Webkit) – Python code – Lucene search engine (java) M. Hearst Faceted Metadata in Search

Metadata Availability • Many collections already have rich metadata associated with them. • Automated methods are improving. • This tool may be helpful for resolving metadata creation wars. M. Hearst Faceted Metadata in Search

Summary • Usability studies done on 3 collections: – Recipes: 13, 000 items – Architecture Images: 40, 000 items – Fine Arts Images: 35, 000 items • Conclusions: – Users like and are successful with the dynamic faceted hierarchical metadata, especially for browsing tasks – Very positive results, in contrast with studies on earlier iterations – Note: it seems you have to care about the contents of the collection to like the interface M. Hearst Faceted Metadata in Search

Other Domains • Applying this to – Text • Tobacco Documents Archives • Medline biomedical texts – Products/Catalogs • Don’t have a collection; would like one M. Hearst Faceted Metadata in Search

Future Work • What about information visualization? • How to integrate with relevance feedback (more like this)? • How to incorporate user preferences and past behavior? • How to combine facets to reflect tasks? M. Hearst Faceted Metadata in Search

Try the Demo: flamenco. berkeley. edu Thanks to: Andrea Sahli Rashmi Sinha NSF CAREER Grant IIS-9984741 IBM Faculty Fellowship 65