Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464129
J. Castro, E. Meneses
Automatic segmentation and classification methods for bioacoustic signals enable real-time monitoring, population estimation, as well as other important tasks for the conservation, management, and study of wildlife. These methods normally require a filter or a denoising strategy to enhance relevant information in the input signal and avoid false positive detections. This denoising stage is usually the performance bottleneck of such methods. In this paper, we parallelize a denoising algorithm for tonal bioacoustic signals using mainly OpenACC directives. The implemented program was executed in both multicore and GPU architectures. The proposed parallelized algorithm achieves a higher speedup on GPU than CPU, leading to a 10.67 speedup compared to the original sequential algorithm in C++.
{"title":"Parallelization of a Denoising Algorithm for Tonal Bioacoustic Signals Using OpenACC Directives","authors":"J. Castro, E. Meneses","doi":"10.1109/IWOBI.2018.8464129","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464129","url":null,"abstract":"Automatic segmentation and classification methods for bioacoustic signals enable real-time monitoring, population estimation, as well as other important tasks for the conservation, management, and study of wildlife. These methods normally require a filter or a denoising strategy to enhance relevant information in the input signal and avoid false positive detections. This denoising stage is usually the performance bottleneck of such methods. In this paper, we parallelize a denoising algorithm for tonal bioacoustic signals using mainly OpenACC directives. The implemented program was executed in both multicore and GPU architectures. The proposed parallelized algorithm achieves a higher speedup on GPU than CPU, leading to a 10.67 speedup compared to the original sequential algorithm in C++.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116343985","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464200
L. Brenes-Guillén, Man-Sai Acón-Chan, R. Mora-Rodríguez, Laura Monturiol-Gross
Clostridium perfringens phospholipase C (CpPLC) plays a key role in the pathogenesis of gas gangrene, an acute and life-threatening infection in humans. CpPLC leads to the unregulated production of second messengers, such as calcium, and activates important signaling pathways like PKC, MEK, and NFkB, inducing ROS production, that leads towards cellular damage. We propose a systems biology approach to identify potential microRNAs which could be used as tools to explore the role of calcium channels in the mechanism of action of CpPLC, and eventually as therapeutic agents that render cells less sensitive to this toxin. Results show that miR19B1 and miR449B are interesting candidate targets which perturbations may affect the expression of genes involved in calcium transport. This approach may be useful to elucidate calcium molecular interactions involved in the CpPLC mechanism of action.
{"title":"A Systems Biology Approach to Investigate Control Targets of Intracellular Calcium Transients Induced by Clostridium Perfringens Phospholipase C","authors":"L. Brenes-Guillén, Man-Sai Acón-Chan, R. Mora-Rodríguez, Laura Monturiol-Gross","doi":"10.1109/IWOBI.2018.8464200","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464200","url":null,"abstract":"Clostridium perfringens phospholipase C (CpPLC) plays a key role in the pathogenesis of gas gangrene, an acute and life-threatening infection in humans. CpPLC leads to the unregulated production of second messengers, such as calcium, and activates important signaling pathways like PKC, MEK, and NFkB, inducing ROS production, that leads towards cellular damage. We propose a systems biology approach to identify potential microRNAs which could be used as tools to explore the role of calcium channels in the mechanism of action of CpPLC, and eventually as therapeutic agents that render cells less sensitive to this toxin. Results show that miR19B1 and miR449B are interesting candidate targets which perturbations may affect the expression of genes involved in calcium transport. This approach may be useful to elucidate calcium molecular interactions involved in the CpPLC mechanism of action.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127025000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464194
Carlos Gamboa-Venegas, Esteban Meneses
Finding the genome of new species remains as one of the most crucial tasks in molecular biology. To achieve that end, de novo sequence assembly feeds from the vast amount of data provided by Next-Generation Sequencing technology. Therefore, genome assemblers demand a high amount of computational resources, and parallel implementations of those assemblers are readily available. This paper presents a comparison of three well-known de novo genome assemblers: Velvet, ABySS and SOAPdenovo, all of them using de Bruijn graphs and having a parallel implementation. We based our analysis on parallel execution time, scalability, quality of assembly, and sensitivity to the choice of a critical parameter (k- mer size). We found one of the tools clearly stands out for providing faster execution time and better quality in the output. Also, all assemblers are mildly sensitive to the choice of k-mer size and they all show limited scalability. We expect the findings of this paper provide a guide to the development of new algorithms and tools for scalable parallel genome sequence assemblers.
{"title":"Comparative Analysis of de Bruijn Graph Parallel Genome Assemblers","authors":"Carlos Gamboa-Venegas, Esteban Meneses","doi":"10.1109/IWOBI.2018.8464194","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464194","url":null,"abstract":"Finding the genome of new species remains as one of the most crucial tasks in molecular biology. To achieve that end, de novo sequence assembly feeds from the vast amount of data provided by Next-Generation Sequencing technology. Therefore, genome assemblers demand a high amount of computational resources, and parallel implementations of those assemblers are readily available. This paper presents a comparison of three well-known de novo genome assemblers: Velvet, ABySS and SOAPdenovo, all of them using de Bruijn graphs and having a parallel implementation. We based our analysis on parallel execution time, scalability, quality of assembly, and sensitivity to the choice of a critical parameter (k- mer size). We found one of the tools clearly stands out for providing faster execution time and better quality in the output. Also, all assemblers are mildly sensitive to the choice of k-mer size and they all show limited scalability. We expect the findings of this paper provide a guide to the development of new algorithms and tools for scalable parallel genome sequence assemblers.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116551136","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464207
K. L. D. Ipiña, N. Barroso, P. Calvo, A. Ezeiza, Carmen Hernández
This paper presents a system for the management of audio information based on semantic knowledge in complex environments. These environments are defined by: the limited resources (audio corpora); the poor quality of the signal (internet radio); the trilingual context (Basque, French and Spanish); the cross-lingual elements between the languages; and the critical under-resourced situation of the Basque minority language. Evolutionary methods have been applied to fulfill the specifications based on: folksonomies, automatic configuration selection, and collaborative networks. The resulting system includes a multilingual ASR engine that is able to tackle under-resourced languages with a very small audio corpus.
{"title":"A Multilingual System Based on Semantic Knowledge for Complex Environments","authors":"K. L. D. Ipiña, N. Barroso, P. Calvo, A. Ezeiza, Carmen Hernández","doi":"10.1109/IWOBI.2018.8464207","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464207","url":null,"abstract":"This paper presents a system for the management of audio information based on semantic knowledge in complex environments. These environments are defined by: the limited resources (audio corpora); the poor quality of the signal (internet radio); the trilingual context (Basque, French and Spanish); the cross-lingual elements between the languages; and the critical under-resourced situation of the Basque minority language. Evolutionary methods have been applied to fulfill the specifications based on: folksonomies, automatic configuration selection, and collaborative networks. The resulting system includes a multilingual ASR engine that is able to tackle under-resourced languages with a very small audio corpus.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123695527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464215
J. Krizaj, Ž. Emeršič, S. Dobrišek, P. Peer, Vitomir Štruc
A novel method for automatic facial landmark localization is presented. The method builds on the supervised descent framework, which was shown to successfully localize landmarks in the presence of large expression variations and mild occlusions, but struggles when localizing landmarks on faces with large pose variations. We propose an extension of the supervised descent framework that trains multiple descent maps and results in increased robustness to pose variations. The performance of the proposed method is demonstrated on the Bosphorus, the FRGC and the UND data sets for the problem of facial landmark localization from 3D data. Our experimental results show that the proposed method exhibits increased robustness to pose variations, while retaining high performance in the case of expression and occlusion variations.
{"title":"Localization of Facial Landmarks in Depth Images Using Gated Multiple Ridge Descent","authors":"J. Krizaj, Ž. Emeršič, S. Dobrišek, P. Peer, Vitomir Štruc","doi":"10.1109/IWOBI.2018.8464215","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464215","url":null,"abstract":"A novel method for automatic facial landmark localization is presented. The method builds on the supervised descent framework, which was shown to successfully localize landmarks in the presence of large expression variations and mild occlusions, but struggles when localizing landmarks on faces with large pose variations. We propose an extension of the supervised descent framework that trains multiple descent maps and results in increased robustness to pose variations. The performance of the proposed method is demonstrated on the Bosphorus, the FRGC and the UND data sets for the problem of facial landmark localization from 3D data. Our experimental results show that the proposed method exhibits increased robustness to pose variations, while retaining high performance in the case of expression and occlusion variations.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125982733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464138
Ž. Emeršič, Nil Oleart Playà, Vitomir Štruc, P. Peer
Automatic ear recognition is gaining popularity within the research community due to numerous desirable properties, such as high recognition performance, the possibility of capturing ear images at a distance and in a covert manner, etc. Despite this popularity and the corresponding research effort that is being directed towards ear recognition technology, open problems still remain. One of the most important issues stopping ear recognition systems from being widely available are ear occlusions and accessories. Ear accessories not only mask biometric features and by this reduce the overall recognition performance, but also introduce new non-biometric features that can be exploited for spoofing purposes. Ignoring ear accessories during recognition can, therefore, present a security threat to ear recognition and also adversely affect performance. Despite the importance of this topic there has been, to the best of our knowledge, no ear recognition studies that would address these problems. In this work we try to close this gap and study the impact of ear accessories on the recognition performance of several state-of-the-art ear recognition techniques. We consider ear accessories as a tool for spoofing attacks and show that CNN-based recognition approaches are more susceptible to spoofing attacks than traditional descriptor-based approaches. Furthermore, we demonstrate that using inpainting techniques or average coloring can mitigate the problems caused by ear accessories and slightly outperforms (standard) black color to mask ear accessories.
{"title":"Towards Accessories-Aware Ear Recognition","authors":"Ž. Emeršič, Nil Oleart Playà, Vitomir Štruc, P. Peer","doi":"10.1109/IWOBI.2018.8464138","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464138","url":null,"abstract":"Automatic ear recognition is gaining popularity within the research community due to numerous desirable properties, such as high recognition performance, the possibility of capturing ear images at a distance and in a covert manner, etc. Despite this popularity and the corresponding research effort that is being directed towards ear recognition technology, open problems still remain. One of the most important issues stopping ear recognition systems from being widely available are ear occlusions and accessories. Ear accessories not only mask biometric features and by this reduce the overall recognition performance, but also introduce new non-biometric features that can be exploited for spoofing purposes. Ignoring ear accessories during recognition can, therefore, present a security threat to ear recognition and also adversely affect performance. Despite the importance of this topic there has been, to the best of our knowledge, no ear recognition studies that would address these problems. In this work we try to close this gap and study the impact of ear accessories on the recognition performance of several state-of-the-art ear recognition techniques. We consider ear accessories as a tool for spoofing attacks and show that CNN-based recognition approaches are more susceptible to spoofing attacks than traditional descriptor-based approaches. Furthermore, we demonstrate that using inpainting techniques or average coloring can mitigate the problems caused by ear accessories and slightly outperforms (standard) black color to mask ear accessories.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125605953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464212
Esteban Arias-Méndez, Alonso Montero-Marín, Danny Chaves-Chaves, F. Torres-Rojas
Comparing two graphs is a computationally difficult task [9], [8]. After a work by E. Arias-Mendez and F. Torres-Rojas [7] about the correlation of metabolic pathways with two new proposed approaches to simplify the comparison of its associated graph representation, we extended this work to general graph structures as a simple way to compare them. The approach presented here is an extension of those algorithms to general graphs. The first algorithm proposed looks to transform the comparing graphs into linear sequences, to be analyzed using sequence-alignment tools from bioinformatics and get a numeric score as its value of similitude. The second proposed algorithm consists of the search of equal connected nodes between 2 graphs to eliminate then on both structures, only leaving the differences, as heuristic for comparison. These algorithms were developed as a low-cost process to correlate metabolic pathways showing good results; the suggestion is to use this information as a previous analysis to a deeper, more expensive, comparing tools use. Here we review the extension of this work as an application to a more general graph data structure. These methods have shown to be an effective way to treat the problem as listed in the results section.
{"title":"Simple Graph Comparison Inspired on Metabolic Pathway Correlation","authors":"Esteban Arias-Méndez, Alonso Montero-Marín, Danny Chaves-Chaves, F. Torres-Rojas","doi":"10.1109/IWOBI.2018.8464212","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464212","url":null,"abstract":"Comparing two graphs is a computationally difficult task [9], [8]. After a work by E. Arias-Mendez and F. Torres-Rojas [7] about the correlation of metabolic pathways with two new proposed approaches to simplify the comparison of its associated graph representation, we extended this work to general graph structures as a simple way to compare them. The approach presented here is an extension of those algorithms to general graphs. The first algorithm proposed looks to transform the comparing graphs into linear sequences, to be analyzed using sequence-alignment tools from bioinformatics and get a numeric score as its value of similitude. The second proposed algorithm consists of the search of equal connected nodes between 2 graphs to eliminate then on both structures, only leaving the differences, as heuristic for comparison. These algorithms were developed as a low-cost process to correlate metabolic pathways showing good results; the suggestion is to use this information as a previous analysis to a deeper, more expensive, comparing tools use. Here we review the extension of this work as an application to a more general graph data structure. These methods have shown to be an effective way to treat the problem as listed in the results section.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133617614","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464210
Man-Sai Acón-Chan, Francisco Siles Canales, R. Mora-Rodríguez
In a previous work we developed a biocomputational platform to automatically construct mathematical models of microRNA-transcription factor (TF) regulatory network to study gene dosage compensation. We hypothesized that gene dosage compensation maintains the viability of tumoral cells with high levels of chromosomal alterations by down regulating the expression of a core group of compensated genes. Here we report that those compensated candidates form clusters according to their chromosomal locations and they share the connections of a large-scale network of putative interactions with miRNAs and TFs. Although initial models of feed-forward loops do not show gene dosage compensation, we simplified the model by including only sensor loops of experimentally validated interactions. A sensor loop allows a gene to sense and react to fluctuations of its own expression. The resulting model shows gene dosage compensation for MYC and STAT3 but the genetic regulation of these two genes can be extended to co-regulate the expression of other compensated genes. These results contribute to the identification of a network of gene dosage compensation, which manipulation of specific nodes has the potential to become a novel approach to specifically target aneuploid cancer cells.
{"title":"Modeling Gene Dosage Compensation Mediated by Sensor Loops in Large-Scale Mathematical Models of Microrna-Transcription Factor Networks","authors":"Man-Sai Acón-Chan, Francisco Siles Canales, R. Mora-Rodríguez","doi":"10.1109/IWOBI.2018.8464210","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464210","url":null,"abstract":"In a previous work we developed a biocomputational platform to automatically construct mathematical models of microRNA-transcription factor (TF) regulatory network to study gene dosage compensation. We hypothesized that gene dosage compensation maintains the viability of tumoral cells with high levels of chromosomal alterations by down regulating the expression of a core group of compensated genes. Here we report that those compensated candidates form clusters according to their chromosomal locations and they share the connections of a large-scale network of putative interactions with miRNAs and TFs. Although initial models of feed-forward loops do not show gene dosage compensation, we simplified the model by including only sensor loops of experimentally validated interactions. A sensor loop allows a gene to sense and react to fluctuations of its own expression. The resulting model shows gene dosage compensation for MYC and STAT3 but the genetic regulation of these two genes can be extended to co-regulate the expression of other compensated genes. These results contribute to the identification of a network of gene dosage compensation, which manipulation of specific nodes has the potential to become a novel approach to specifically target aneuploid cancer cells.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130707352","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464201
J. Becerra, R. Duro, J. Monroy, F. Bellas
Perceptual classes allow cognitive systems to organize the continuous perceptual space they are usually faced with into discrete categories so that planning and long-term prospection can take place. However, for the correct operation of the system it is quite important that the perceptual classes are precisely delimited in order to prevent classification errors that would lead to planning and prediction errors. In this paper we propose a simple but effective mechanism that makes use of the prospective capabilities of a cognitive architecture during a dreaming phase in order to allow it to generate hypotheses on the delimitation of perceptual classes. These hypotheses are later tested as the system interacts with the world when in an awake state.
{"title":"A Dreaming Approach to Perceptual Class Delimitation Within the DREAM Architecture","authors":"J. Becerra, R. Duro, J. Monroy, F. Bellas","doi":"10.1109/IWOBI.2018.8464201","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464201","url":null,"abstract":"Perceptual classes allow cognitive systems to organize the continuous perceptual space they are usually faced with into discrete categories so that planning and long-term prospection can take place. However, for the correct operation of the system it is quite important that the perceptual classes are precisely delimited in order to prevent classification errors that would lead to planning and prediction errors. In this paper we propose a simple but effective mechanism that makes use of the prospective capabilities of a cognitive architecture during a dreaming phase in order to allow it to generate hypotheses on the delimitation of perceptual classes. These hypotheses are later tested as the system interacts with the world when in an awake state.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122663265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/IWOBI.2018.8464216
Geovanni Figueroa-Mata, Erick Mata-Montero, Juan Carlos Valverde-Otarola, Dagoberto Arias-Aguilar
Forest species identification is critical to scientifically support many environmental, commercial, forensic, archaeological, and paleontological actividades. Therefore, it is very important to develop fast and accurate identification systems. We present a deep CNN for automated forest species identification based on macroscopic images of wood cuts. We first implement and study a modified version of the LeNet convolutional network, which is trained from scratch with a database of macroscopic images of 41 forest species of the Brazilian flora. With this network we achieve a top-1 accuracy of 93.6%. Additionally, we fine-tune the Resnet50 model with pre-trained weights on Imagenet to reach a top-1 accuracy of 98.03%, which improves previous published results of research on the same image database.
{"title":"Using Deep Convolutional Networks for Species Identification of Xylotheque Samples","authors":"Geovanni Figueroa-Mata, Erick Mata-Montero, Juan Carlos Valverde-Otarola, Dagoberto Arias-Aguilar","doi":"10.1109/IWOBI.2018.8464216","DOIUrl":"https://doi.org/10.1109/IWOBI.2018.8464216","url":null,"abstract":"Forest species identification is critical to scientifically support many environmental, commercial, forensic, archaeological, and paleontological actividades. Therefore, it is very important to develop fast and accurate identification systems. We present a deep CNN for automated forest species identification based on macroscopic images of wood cuts. We first implement and study a modified version of the LeNet convolutional network, which is trained from scratch with a database of macroscopic images of 41 forest species of the Brazilian flora. With this network we achieve a top-1 accuracy of 93.6%. Additionally, we fine-tune the Resnet50 model with pre-trained weights on Imagenet to reach a top-1 accuracy of 98.03%, which improves previous published results of research on the same image database.","PeriodicalId":127078,"journal":{"name":"2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124777905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：