Multimodal Interactive handwritten text transcription / / Veronica Romero, Alejandra Hector Toselli, Enrique Vidal

• Record Nr.: UNINA9910807924103321
• Autore: Romer Veronica
• Pubbl/distr/stampa: Singapore ; ; Hackensack, NJ, : World Scientific Pub, c2012
• ISBN: 1-283-59373-4; 9786613906182; 981-4390-34-8
• Edizione: [1st ed.]
• Descrizione fisica: 1 online resource (180 p.)
• Collana: Series in machine perception and artificial intelligence ; ; v. 80
• Altri autori (Persone): ToselliAlejandro Hector; VidalEnrique
• Disciplina: 006.425
• Soggetti: Writing - Data processing; Multimodal user interfaces (Computer systems); Human-computer interaction
• Lingua di pubblicazione: Inglese
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Note generali: Description based upon print version of record.
• Nota di bibliografia: Includes bibliographical reference (p. 155-164) and index.
• Nota di contenuto: Contents; Preface; 1. Preliminaries; 1.1 Introduction; 1.2 State of the Art; 1.2.1 Optical Character Recognition; 1.2.2 Handwritten Text Recognition; 1.3 Formal Background; 1.3.1 Hidden Markov Models; Continuous HMM; Basic algorithms for HMMs; The Decoding Problem and the Viterbi Algorithm; The Learning Problem and the Baum-Welch Algorithm; 1.3.2 Language models: N-grams; n-grams modelled by a stochastic finite state automaton; 1.3.3 Interactive Pattern Recognition; 1.3.4 Word-graphs; 1.4 Assessing Computer Assisted Transcription of Handwritten Text Images; 2. Corpora; 2.1 Introduction; 2.2 CS2.3 ODEC; 2.4 IAMDB; 2.5 UNIPEN; 3. Handwritten Text Recognition; 3.1 Introduction; 3.2 Off-line Handwritten Text Recognition; 3.2.1 Preprocessing; 3.2.2 Feature Extraction; 3.2.3 Recognition; 3.2.4 Experimental Framework; 3.2.5 Meta-parameter Adjustment Experiments; 3.2.6 Discussion of Results; 3.3 On-line Handwritten Text Recognition; 3.3.1 Preprocessing; 3.3.2 Feature Extraction; 3.3.3 Recognition; 3.3.4 Experimental Framework; 3.3.5 Results; 3.4 Summary and Conclusions; 4. Computer Assisted Transcription of Handwritten Text Images; 4.1 Introduction; 4.2 Formal ; Framework; 4.3 Adapting the Language Model4.4 Searching; 4.4.1 Direct Viterbi-based Approach; 4.4.2 Word-graph based Approach; 4.4.2.1 Error-correction parsing; 4.5 Increasing Interaction Ergonomy; 4.5.1 Language Modelling and Search; 4.6 Interacting at the Character Level; 4.6.1 Language Modelling and Search; 4.7 Experimental Framework; 4.7.1 Assessment Measures; 4.7.2 Parameters and Meta-Parameters; 4.8 Results; Direct Viterbi-based approach; Word-graph based approach; Using Pointer Actions in the CATTI interaction process (PA- CATTI); CATTI at the character level; 4.9 Conclusions and Future Work; 5. Multimodal Computer Assisted Transcription of Handwritten Text Images5.1 Introduction; 5.2 Formal Framework; 5.3 Adapting the Language Model; 5.4 Searching; 5.5 Experimental Framework; 5.5.1 Corpora; 5.5.2 Assessment Measures; 5.6 Results; 5.7 Conclusions; 6. A Web-based Demonstrator of Interactive Multimodal Transcription; 6.1 Introduction; 6.2 User Interaction Protocol; 6.3 System Description; 6.3.1 Application Programming Interface; 6.3.2 MM-CATTI Server; 6.3.3 Web Interface; 6.3.4 Electronic Pen or Touchscreen Interaction; 6.3.5 Keyboard and Mouse Interaction; 6.4 Evaluation; 6.4.1 Assessment Measures6.4.2 Corpus; 6.4.3 Participants; 6.4.4 Apparatus; 6.4.5 Procedure; 6.4.6 Design; 6.5 Results and Discussion; 6.5.1 Quantitative Analysis; 6.5.1.1 Analysis of Time; 6.5.1.2 Analysis of rWER; 6.5.1.3 Analysis of WSR; 6.5.2 Qualitative Analysis; 6.5.3 Correlation Analysis; 6.5.3.1 Correlation between trials; 6.5.3.2 Correlation between metrics; 6.5.4 Limitations of the Study; 6.6 Conclusions; 7. Conclusions and Outlook; 7.1 Conclusions; 7.2 Outlook; Acknowledgements; Appendix A Symbols and Acronyms; A.1 Symbols; A.2 Acronyms; Bibliography; Index
• Sommario/riassunto: This book presents an interactive multimodal approach for efficient transcription of handwritten text images. This approach, rather than full automation, assists the expert in the recognition and transcription process.Until now, handwritten text recognition (HTR) systems are far from being perfect and heavy human intervention is often required to check and correct the results of such systems. The interactive scenario studied in this book combines the efficiency of automatic handwriting recognition systems with the accuracy of the experts, leading to a cost-effective perfect transcription of th