Evaluation of new approaches to the access of official monitoring results for live bivalves molluscs
Abstract
Abstract Background Live bivalve molluscs, echinoderms, tunicates and marine gastropod are referred in EU food laws, and require member states to implement official controls in classified production areas, with the monitoring and classification of those areas. If, due to contaminant tests results, a production area is closed, any product from there is prohibited to be commercialized. Mobile applications optical character recognition (OCR) functionalities could ease the access to contaminant levels and production area classifications. This study verifies what information is available in live bivalves’ labels, describes an OCR algorithm for those labels and evaluates it. Methods 86 labels were selected from four sale points in Lisbon, and photographed using smartphones. Each label was evaluated by a human to determine what data was available (either required or not). An OCR algorithm was developed and applied on the collected labels and validated against the data extracted by the human analysis. Results The analysis shows that all the labels included the required information, and 63% of the labels included the identifier for the production zone. The label-reading algorithm performs with an accuracy of 79.85% for the individual values. Conclusion High accuracy of the developed label-reading algorithm shows potential for providing instant automatic accessmore »
- Authors:
-
[1]
- INESC-ID Lisboa, Instituto Superior Técnico, Universidade de Lisboa, Lisboa, Portugal
- Tulane University School of Public Health and Tropical Medicine, New Orleans, Louisiana, USA
- Instituto Superior Técnico, Universidade de Lisboa, Lisboa, Portugal
- Publication Date:
- Sponsoring Org.:
- USDOE Office of Nuclear Energy (NE), Fuel Cycle Technologies (NE-5)
- OSTI Identifier:
- 1566906
- Grant/Contract Number:
- UID/CEC/50021/2019
- Resource Type:
- Published Article
- Journal Name:
- European Journal of Public Health
- Additional Journal Information:
- Journal Name: European Journal of Public Health Journal Volume: 29 Journal Issue: 5; Journal ID: ISSN 1101-1262
- Publisher:
- Oxford University Press
- Country of Publication:
- Country unknown/Code not available
- Language:
- English
Citation Formats
da Costa, Manuel Beja, Jurczuk, Magdalena, Marques, Bernardo, and Silva, João Nuno. Evaluation of new approaches to the access of official monitoring results for live bivalves molluscs. Country unknown/Code not available: N. p., 2019.
Web. doi:10.1093/eurpub/ckz018.
da Costa, Manuel Beja, Jurczuk, Magdalena, Marques, Bernardo, & Silva, João Nuno. Evaluation of new approaches to the access of official monitoring results for live bivalves molluscs. Country unknown/Code not available. doi:10.1093/eurpub/ckz018.
da Costa, Manuel Beja, Jurczuk, Magdalena, Marques, Bernardo, and Silva, João Nuno. Wed .
"Evaluation of new approaches to the access of official monitoring results for live bivalves molluscs". Country unknown/Code not available. doi:10.1093/eurpub/ckz018.
@article{osti_1566906,
title = {Evaluation of new approaches to the access of official monitoring results for live bivalves molluscs},
author = {da Costa, Manuel Beja and Jurczuk, Magdalena and Marques, Bernardo and Silva, João Nuno},
abstractNote = {Abstract Background Live bivalve molluscs, echinoderms, tunicates and marine gastropod are referred in EU food laws, and require member states to implement official controls in classified production areas, with the monitoring and classification of those areas. If, due to contaminant tests results, a production area is closed, any product from there is prohibited to be commercialized. Mobile applications optical character recognition (OCR) functionalities could ease the access to contaminant levels and production area classifications. This study verifies what information is available in live bivalves’ labels, describes an OCR algorithm for those labels and evaluates it. Methods 86 labels were selected from four sale points in Lisbon, and photographed using smartphones. Each label was evaluated by a human to determine what data was available (either required or not). An OCR algorithm was developed and applied on the collected labels and validated against the data extracted by the human analysis. Results The analysis shows that all the labels included the required information, and 63% of the labels included the identifier for the production zone. The label-reading algorithm performs with an accuracy of 79.85% for the individual values. Conclusion High accuracy of the developed label-reading algorithm shows potential for providing instant automatic access to the date and production area, but is affected by the variability on the label structure. Although not required by food laws, the majority of the sampled labels included complementary information (classified production area) that will allow access to more precise information about the existing biotoxin tests and analysis results.},
doi = {10.1093/eurpub/ckz018},
journal = {European Journal of Public Health},
number = 5,
volume = 29,
place = {Country unknown/Code not available},
year = {2019},
month = {2}
}
DOI: 10.1093/eurpub/ckz018
Search WorldCat to find libraries that may hold this journal