Multimedia Databases - IfIS - Technische Universität Braunschweig
Transcription
Multimedia Databases - IfIS - Technische Universität Braunschweig
10/27/2009 0. Organizational Issues • Lecture – 22.10.2009 – 04.02.2010 – 15:00-17:15 (3 lecture hours with a break) – Exercises, detours, and home work discussion integrated into lecture Multimedia Databases • 4 Credits • Exams Wolf-Tilo Balke Silviu Homoceanu Institut für Informationssysteme Technische Universität Braunschweig http://www.ifis.cs.tu-bs.de – Oral exam – 50% of exercise points needed to be eligible for the exam Multimedia Databases – Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 0. Organizational Issues 0. Organizational Issues – Castelli/Bergman: Image Databases, Wiley, 2002 • Recommended literature – Schmitt: Ähnlichkeitssuche in Multimedia-Datenbanken, Oldenbourg, 2005 – Khoshafian/Baker: Multimedia and Imaging Databases, Morgan Kaufmann, 1996 – Steinmetz: Multimedia-Technologie: Grundlagen, Komponenten und Systeme, Springer, 1999 Relational Database Systems 1 – Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig – Sometimes: original papers (on our Web page) 3 0. Organizational Issues Relational Database Systems 1 – Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 4 1. Introduction • Course Web page 1. Introduction – http://www.ifis.cs.tu-bs.de/teaching/ws-0910/mmdb 1.1 What are multimedia databases? 1.2 Multimedia database applications 1.3 Evaluation of retrieval techniques – Contains slides, excercises, related papers and a video of the lecture – Any questions? Just drop us an email… Relational Database Systems 1 – Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 2 5 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 6 1 10/27/2009 1.1 Multimedia Databases 1.1 Basic Definitions • What are multimedia databases (MMDB)? • Multimedia – Databases + multimedia = MMDB – The concept of multimedia expresses the integration of different digital media types – The integration is usually performed in a document – Basic media types are text, image, vector graphics, audio and video • Key words: databases and multimedia • We already know databases, so what is multimedia? Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 7 1.1 Data Types Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 8 1.1 Documents • Text • Earliest definition of information retrieval: “Documents are logically interdependent digitally encoded texts“ • Extension to multimedia documents allows the additional integration of other media types as images, audio or video – Text data, Spreadsheets, E-Mail, … • Image – Photos (Bitmaps), Vector graphics, CAD, … • Audio – Speech- and music records, annotations, wave files, MIDI, MP3, … • Video – Dynamical image record, frame-sequences, MPEG, AVI, … Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 9 1.1 Documents 10 1.1 Basic Definitions • Document types • Medium – Media objects are documents which are of only one type (not necessarily text) – Multimedia objects are general documents which allow an arbitrary combination of different types – A medium is a carrier of information in a communication connection – It is independent of the transported information – The used medium can also be changed during information transfer • Multimedia data is transferred through the use of a medium Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 11 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 12 2 10/27/2009 1.1 Medium Example 1.1 Medium Classification • Book • Based on receiver type – Visual/optical medium – Acoustic mediums – Haptical medium – through tactile senses – Olfactory medium – through smell – Gustatory medium – through taste – Communication between author and reader – Independent from content – Hierarchically built on text and images – Reading out loud represents medium change to sound/audio Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig • Based on time – Dynamic – Static 13 1.1 Multimedia Databases • Persistent storage of multimedia data, e.g.: – What multimedia is – And how it is transported (through some medium) – Text documents – Vector graphics, CAD – Images, audio, video • But… why do we need databases? • Content-based retrieval – Most important operations of databases are data storage and data retrieval – Efficient content based search – Standardization of meta-data (e. g., MPEG-7, MPEG-21) 15 1.1 Multimedia Databases 16 1960 – Special retrieval functionality as well as corresponding optimization can be provided in both cases… – But in the second case we also get the general advantages of databases Retrieval procedures for text documents (Information Retrieval) 1970 Relational Databases and SQL 1980 Presence of multimedia objects intensifies Declarative query language Orthogonal combination of the query functionality Query optimization, Index structures Transaction management, recovery ... Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 1.1 Historical Overview • Stand-alone vs. database storage model? • • • • • 14 1.1 Multimedia Databases • We now have seen… Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 1990 SQL-92 introduces BLOBs First Multimedia-Databases 2000 17 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 18 3 10/27/2009 1.1 Commercial Systems 1.1 Requirements • Relational Databases use the data type BLOB (binary large object) • Requirements for multimedia databases (Christodoulakis, 1985) – Uninterpreted data – Retrieval through metadata like e.g., file name, size, author, … – Classical database functionality – Maintenance of unformatted data – Consideration of special storage and presentation devices • Object-relational extensions feature enhanced retrieval functionality – Semantic search – IBM DB2 Extenders, Oracle Cartridges, … – Integration in DB through UDFs, UDTs, Stored Procedures, … Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 19 1.1 Requirements – User interface – how should the user interact with the system? Separate structure from content! – Information extraction – (automatic) generation of content-based features – Storage devices – very large storage capacity, redundancy control and compression – Information retrieval – integration of some extended search functionality – Software architecture – new or extension of existing databases? – Content addressing – identification of the objects through content-based features – Performance – improvements using indexes, optimization, etc. 21 1.1 Retrieval Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 22 1.1 Retrieval • Retrieval means the choice between data objects which can be based on… • Closer look at the search functionality – „Semantic“ search functionality – Orthogonal integration of classical and extended functionality – Search does not directly access the media objects – Extraction, normalization and indexing of contentbased features – Meaningful similarity/distance measures – a SELECT condition (exact match) – or a defined similarity connection (best match) • Retrieval may cover the delivery of the results to the user, too Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 20 1.1 Requirements • To comply with these requirements the following aspects need to be considered Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 23 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 24 4 10/27/2009 1.1 Content-based Retrieval • “Retrieve all images showing a sunset !” 1.1 Schematic View • Usually 2 main steps – Example: image databases Image collection Image analysis and feature extraction Digitization Image database Creating the database Querying the database Image query • What exactly do these images have in common? Image analysis and feature extraction Digitization Similarity search Search result 25 1.1 Detailed View Query Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Query Result Result Query preparation 3. Query preparation 5. Result preparation Query plan & feature values Result data 4. Similarity computation & query processing Feature values Normalization Segmentation Feature extraction Optimization Feature values Query plan Similarity computation Raw & relational data Raw data 2. Extraction of features 26 1.1 More Detailed View Relevance feedback Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Result preparation Medium transformation Format transformation Result data Query processing Feature values MM Database MM-Database Feature index BLOBs/CLOBs MM-Database Relational DB Metadata Profile Structure data Feature extraction Pre-processing 1. Insert into the database Feature recognition Feature preparation Decomposition Normalization Segmentation MM-Objects + relational data MM-Objects Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 27 1.2 Applications 28 1.2 Applications • Lots of multimedia content on the Web • Cameras are everywhere – Social networking e.g., Facebook, MySpace, Hi5, Orkut, etc. – Photo sharing e.g., Flickr, Photobucket, Imeem, Picasa, etc. – Video sharing e.g., YouTube, Megavideo, Metacafe, blip.tv, Liveleak, etc. Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Relational data Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 29 – In London “there are at least 500,000 cameras in the city, and one study showed that in a single day a person could expect to be filmed 300 times” Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 30 5 10/27/2009 1.2 Applications 1.2 Applications • Picasa face recognition Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig • Picasa, face recognition example 31 1.2 Applications • Picasa example 33 1.2 Sample Scenario Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 34 1.2 Sample Scenario • Consider a police investigation of a large-scale drug operation • Possible generated data: • Possible queries – Image query by example: police officer has a photograph and wants to find the identity of the person in the picture – Video data captured by surveillance cameras – Audio data captured – Image data consisting of still photographs taken by investigators – Structured relational data containing background information – Geographic information system data Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 32 1.2 Applications • Picasa, learning phase Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig • Query: “retrieve all images from the image library in which the person appearing in the (currently displayed) photograph appears” – Image query by keywords: police officer wants to examine pictures of “Tony Soprano” • Query: “retrieve all images from the image library in which ‘Tony Soprano’ appears" 35 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 36 6 10/27/2009 1.2 Sample Scenario 1.2 Sample Scenario – Video Query: – Heterogeneous Multimedia Query: • Query: “Find all video segments in which Jerry appears” • By examining the answer of the above query, the police officer hopes to find other people who have previously interacted with the victim Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig • Find all individuals who have been photographed with “Tony Soprano” and who have been convicted of attempted murder in New Jersey and who have recently had electronic fund transfers made into their bank accounts from ABC Corp. 37 1.2 Characteristics 38 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 1.2 Example • Basic difference • Passive static retrieval – Static – Art historical use case • High number of search queries (read access), few modifications of the data state – Dynamic • Often modifications of the data state – Active • Database functionality lead to application operations – Passive • Database reacts only at requests from outside – Standard search • Queries are answered through the use of metadata e.g., Google-image search – Retrieval functionality • Content based search on the multimedia repository e.g., Picasa face recognition Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 39 1.2 Example 40 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 1.2 Example – Possible hit in a multimedia database • Active dynamic retrieval – Wetter warning through evaluation of satellite photos Extraction Typhoon-Warning for the Philippines Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 41 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 42 7 10/27/2009 1.2 Example 1.2 Example • Standard search • Retrieval functionality – Queries are answered through the use of metadata e.g., Google-image search Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig – Content based e.g., Picasa face recognition 43 1.3 Towards Evaluation Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 44 1.3 Evaluating Efficiency • Basic evaluation of retrieval techniques • Characteristic values to measure efficiency are e.g.: – Efficiency of the system • Efficient utilization of system resources • Scalable also over big collections – Memory usage – CPU-time – Number of I/O-Operations – Response time – Effectivity of the retrieval process • High quality of the result • Meaningful usage of the system • Depends on the (Hardware-) environment – Weighting is application specific • Goal: the system should be efficient enough! Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 45 1.3 Evaluating Effectivity 46 1.3 Evaluating Effectivity • Measuring effectivity is more difficult and always depending on the query • Goal: define some query-dependent evaluation measures! – Objective quality metrics – Result evaluation based on the query – Independent from used querying interface and retrieval procedure – Leads to comparability of different systems/algorithms Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 47 • Effectivity can be measured regarding some explicit query – Main focus on evaluating the behavior of the system with respect to a query – Relevance of the result set • But effectivity also needs to consider implicit information needs – Main focus on evaluating the usefulness, usability and user friendliness of the system Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 48 8 10/27/2009 1.3 Relevance 1.3 Usefulness (Pertinence) • To evaluate a retrieval system over some query, each document will be classified binary as relevant or irrelevant with respect to the query • Subjective measure estimating to what degree the information need of the user is satisfied – Difficult to measure (empirical studies) – Questionable instrument for comparing procedures/systems – This classification is performed by “experts” – The response of the system to the query will be compared to this manual classification • Attention: useful documents can be irrelevant when considering the query (serendipity) • In this lecture: explicit query evaluation measures only! • Compare the obtained response with the “ideal” result Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 49 1.3 Involved Sets Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 50 1.3 False Positives • Irrelevant documents, classified as relevant by the system collection searched for (= relevant) – False alarms, false drops, … found (= query result) Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig • Needlessly increase the result set • Usually inevitable (ambiguity) • Can be easily eliminated by the user 51 1.3 False Negatives Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 52 1.3 Remaining Sets • Relevant documents classified by the system as irrelevant • Correct positives (correct alarms) – All documents correctly classified by the system as relevant – False dismissals • Correct negatives (correct dismissals) – All documents correctly classified by the system as irrelevant • Dangerous, since they can not be detected easily by the user – Does the collection contain “better” documents? – False positives are usually not as bad as false negatives Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 53 • All sets are disjunctive and their reunion is the entire document collection Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 54 9 10/27/2009 1.3 Overview Systemevaluation 1.3 Interpretation relevant irrelevant relevant ca fd irrelevant fa cd Userevaluation • {Relevant results} = fd + ca • {Retrieved results} = ca + fa collection fd ca searched for fa found cd Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 55 1.3 Precision Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 56 1.3 Recall • Precision measures the ratio of correctly returned documents relative to all returned documents • Recall measures the ratio of correctly returned documents relative to all relevant documents – R = ca / (ca + fd) – P = ca / (ca + fa) • Value between [0, 1] (1 representing the best value) • High number of false drops mean worse results • Value between [0, 1] (1 representing the best value) • High number of false alarms mean worse results Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 57 1.3 Precision-Recall Analysis • Both measures only make sense, if considered at the same time 58 1.3 Actual Evaluation • Alarms (returned elements) can be easily divided in ca and fa – E.g., get perfect recall by simply returning all documents, but then the precision is extremely low… • Can be balanced by tuning the system – E.g., smaller result sets lead to better precision rates at the cost of recall – Precision is easy to calculate • Dismissals (not returned elements) are not so trivial to divide in cd und fd, because the entire collection has to be classified – Recall is difficult to calculate • Standardized Benchmarks • Usually the average precision-recall of more queries is considered (macro evaluation) Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig – Provided connections and queries – Annotated result sets 59 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 60 10 10/27/2009 1.3 TREC 1.3 Example • Text REtrieval Conference • De-Facto-Standard since 1992 • Establish average precision for 11 fixed recall points (0; 0,1; 0,2; …; 1) according to defined procedures (trec_eval) • Different tracks, extended also for video data, ´Web retrieval and Question Answering • Other initiatives: e.g., CLEF (cross-language retrieval) or INEX (XML-Documents) Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig fa ca fd cd P R Q1 8 2 6 4 0,2 0,25 Q2 2 8 2 8 0,8 0,8 0,5 0,525 Average 61 Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 1.3 Representation 62 Next lecture • Precision-Recall-Graphs • • • • Average precision of the system 3 at a recall-level of 0,2 System 1 Query Retrieval of images by color Introduction to color spaces Color histograms Matching Which system is the best? System 2 System 3 What is more important: recall or precision? Multimedia Databases– Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 63 Multimedia Databases – Wolf-Tilo Balke – Institut für Informationssysteme – TU Braunschweig 64 11