Pub Date : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8559068
Pushyami Kaveti, Hanumant Singh
Fisheries independent data one of the most important sources of information for fish stock assessments. Historically these data have been collected by a tools such as bottom trawls which are not effective or desirable in rocky or protected areas, In the last decade we have made significant progress in terms of using robotic platforms[1] [2] to collect optical imagery to assess fish stocks. We now routinely collect hundreds of thousands of images over a single research expedition. Fisheries biologists are overwhelmed by the large datasets that are being collected. In this paper we look at Convolutional Neural Networks [3] [4] as a mechanism to automatically detect and classify fish in underwater imagery. We present the results of analyzing a large dataset of underwater imagery comprising 10,000 images taken by the Seabed Autonomous Underwater Vehicle. The data is diverse - across different habitats, it exhibits no rotational symmetry, has large shadows compared to the organisms under consideration and also has large occlusions and objects that are small and not centered compared to the overall field of view. Despite these serious differences compared to land based image datasets we show that our segmentation and classification results are similar to state of the art efforts associated with land based applications.
{"title":"Towards Automated Fish Detection Using Convolutional Neural Networks","authors":"Pushyami Kaveti, Hanumant Singh","doi":"10.1109/OCEANSKOBE.2018.8559068","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8559068","url":null,"abstract":"Fisheries independent data one of the most important sources of information for fish stock assessments. Historically these data have been collected by a tools such as bottom trawls which are not effective or desirable in rocky or protected areas, In the last decade we have made significant progress in terms of using robotic platforms[1] [2] to collect optical imagery to assess fish stocks. We now routinely collect hundreds of thousands of images over a single research expedition. Fisheries biologists are overwhelmed by the large datasets that are being collected. In this paper we look at Convolutional Neural Networks [3] [4] as a mechanism to automatically detect and classify fish in underwater imagery. We present the results of analyzing a large dataset of underwater imagery comprising 10,000 images taken by the Seabed Autonomous Underwater Vehicle. The data is diverse - across different habitats, it exhibits no rotational symmetry, has large shadows compared to the organisms under consideration and also has large occlusions and objects that are small and not centered compared to the overall field of view. Despite these serious differences compared to land based image datasets we show that our segmentation and classification results are similar to state of the art efforts associated with land based applications.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"135 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123494725","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8559307
Tzu‐Hao Lin, Yu Tsao
Information on the dynamics of the deep-sea ecosystem is essential for conservation management. The marine soundscape has been considered as an acoustical sensing platform to investigate geophysical events, marine biodiversity, and human activities. However, analysis of the marine soundscape remains difficult because of the influence of simultaneous sound sources. In this study, we integrated machine learning-based information retrieval techniques to analyze the variability of the marine soundscape off northeastern Taiwan. A long-term spectral average was employed to visualize the long-duration recordings of the Marine Cable Hosted Observatory (MACHO). Biotic and abiotic soundscape components were separated by applying periodicity-coded nonnegative matrix factorization. Finally, various acoustic events were identified using k-means clustering. Our results show that the MACHO recordings of June 2012 contain multiple sound sources. Cetacean vocalizations, an unidentified biological chorus, environmental noise, and system noise can be accurately separated without an audio recognition database. Cetacean vocalizations were primarily detected at night, which is consistent with the detection results of two rule-based detectors. The unidentified biological chorus, ranging between 2 and 3 kHz, was primarily recorded between 7 p.m. and midnight during the studied period. On the basis of source separation, more acoustic events can be identified in the clustering result. The proposed information retrieval techniques effectively reduce the difficulty in the analysis of marine soundscape. The unsupervised approach of source separation and clustering can improve the investigation regarding the temporal behavior and spectral characteristics of different sound sources. Based on the findings in the present study, we believe that variability of the deep-sea ecosystem can be efficiently investigated by combining the soundscape information retrieval techniques and cabled hydrophone networks in the future.
{"title":"Listening to the Deep: Exploring Marine Soundscape Variability by Information Retrieval Techniques","authors":"Tzu‐Hao Lin, Yu Tsao","doi":"10.1109/OCEANSKOBE.2018.8559307","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8559307","url":null,"abstract":"Information on the dynamics of the deep-sea ecosystem is essential for conservation management. The marine soundscape has been considered as an acoustical sensing platform to investigate geophysical events, marine biodiversity, and human activities. However, analysis of the marine soundscape remains difficult because of the influence of simultaneous sound sources. In this study, we integrated machine learning-based information retrieval techniques to analyze the variability of the marine soundscape off northeastern Taiwan. A long-term spectral average was employed to visualize the long-duration recordings of the Marine Cable Hosted Observatory (MACHO). Biotic and abiotic soundscape components were separated by applying periodicity-coded nonnegative matrix factorization. Finally, various acoustic events were identified using k-means clustering. Our results show that the MACHO recordings of June 2012 contain multiple sound sources. Cetacean vocalizations, an unidentified biological chorus, environmental noise, and system noise can be accurately separated without an audio recognition database. Cetacean vocalizations were primarily detected at night, which is consistent with the detection results of two rule-based detectors. The unidentified biological chorus, ranging between 2 and 3 kHz, was primarily recorded between 7 p.m. and midnight during the studied period. On the basis of source separation, more acoustic events can be identified in the clustering result. The proposed information retrieval techniques effectively reduce the difficulty in the analysis of marine soundscape. The unsupervised approach of source separation and clustering can improve the investigation regarding the temporal behavior and spectral characteristics of different sound sources. Based on the findings in the present study, we believe that variability of the deep-sea ecosystem can be efficiently investigated by combining the soundscape information retrieval techniques and cabled hydrophone networks in the future.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"185 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131612066","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8558823
M. Mochizuki, K. Uehira, T. Kanazawa, T. Kunugi, K. Shiomi, S. Aoi, T. Matsumoto, N. Takahashi, N. Chikasada, Takeshi Nakamura, S. Sekiguchi, M. Shinohara, Tomoaki Yamada
S-net is the online and real-time seafloor observation network of 150 observatories for earthquakes and tsunamis along the Japan Trench. It covers the focal region of the 2011 off the Pacific coast of Tohoku earthquake and its vicinity regions. It was established to enhance reliability of early warnings of tsunami and earthquake after the occurrence of the earthquake. Full-scale operation of the S-net has started since April 2017. The data from the 150 seafloor observatories are being transferred to the data center at NIED on a real-time basis, and then verification of data integrity are being carried out. Obtained data reveal that the S-net can observe and monitor seismic and tsunami phenomena that had been never covered only by land-based observations.
{"title":"S-Net project: Performance of a Large-Scale Seafloor Observation Network for Preventing and Reducing Seismic and Tsunami Disasters","authors":"M. Mochizuki, K. Uehira, T. Kanazawa, T. Kunugi, K. Shiomi, S. Aoi, T. Matsumoto, N. Takahashi, N. Chikasada, Takeshi Nakamura, S. Sekiguchi, M. Shinohara, Tomoaki Yamada","doi":"10.1109/OCEANSKOBE.2018.8558823","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8558823","url":null,"abstract":"S-net is the online and real-time seafloor observation network of 150 observatories for earthquakes and tsunamis along the Japan Trench. It covers the focal region of the 2011 off the Pacific coast of Tohoku earthquake and its vicinity regions. It was established to enhance reliability of early warnings of tsunami and earthquake after the occurrence of the earthquake. Full-scale operation of the S-net has started since April 2017. The data from the 150 seafloor observatories are being transferred to the data center at NIED on a real-time basis, and then verification of data integrity are being carried out. Obtained data reveal that the S-net can observe and monitor seismic and tsunami phenomena that had been never covered only by land-based observations.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125510553","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8559368
S. Flögel, I. Ahrns, C. Nuber, M. Hildebrandt, A. Duda, J. Schwendner, D. Wilde
The exploration of space and deep-sea environments faces significant similarities. As in space, the exploration and utilization of the deep sea is performed under extreme environmental conditions. Recently, deep sea systems are becoming increasingly autonomous, resulting in challenges that are similar to autonomous space systems such as limited energy supply, communication, as well as navigation system control and failure handling. The analogies between autonomous robotic space and deep-sea technologies motivated the German Helmholtz Association to setup the joint research program ROBEX (Robotic Exploration of Extreme environments). In this research program, scientists and engineers from both domains cooperated to find solutions to similar challenges and to mutually benefit from each other's technologies and capabilities. ROBEX consisted of a consortium of German maritime and space research institutions and was funded from 2012–2017. Within the deep-sea crawler project MANSIO-VIATOR, a consortium of marine and space-related institutes developed a new underwater system uniting the advantages of a fixed sea-floor observatory harboring a mobile crawler component to map and monitor large areas on the seafloor.
{"title":"A New Deep-Sea Crawler System - MANSIO-VIATOR","authors":"S. Flögel, I. Ahrns, C. Nuber, M. Hildebrandt, A. Duda, J. Schwendner, D. Wilde","doi":"10.1109/OCEANSKOBE.2018.8559368","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8559368","url":null,"abstract":"The exploration of space and deep-sea environments faces significant similarities. As in space, the exploration and utilization of the deep sea is performed under extreme environmental conditions. Recently, deep sea systems are becoming increasingly autonomous, resulting in challenges that are similar to autonomous space systems such as limited energy supply, communication, as well as navigation system control and failure handling. The analogies between autonomous robotic space and deep-sea technologies motivated the German Helmholtz Association to setup the joint research program ROBEX (Robotic Exploration of Extreme environments). In this research program, scientists and engineers from both domains cooperated to find solutions to similar challenges and to mutually benefit from each other's technologies and capabilities. ROBEX consisted of a consortium of German maritime and space research institutions and was funded from 2012–2017. Within the deep-sea crawler project MANSIO-VIATOR, a consortium of marine and space-related institutes developed a new underwater system uniting the advantages of a fixed sea-floor observatory harboring a mobile crawler component to map and monitor large areas on the seafloor.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132871393","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 : 2018-05-28DOI: 10.1109/oceanskobe.2018.8559136
U. Nielsen, Astrid H. Brodtkorb
This paper presents a study focused on a newly developed procedure for wave spectrum estimation using wave-induced motion recordings from a ship. The particular procedure stands out from other existing, similar ship motion-based procedures by its computational efficiency and - at the same time - providing accurate estimates of the on-site wave conditions. In the paper, the procedure is applied to full-scale experimental data obtained from dedicated sea trial runs. The results show favorable agreement with corresponding wave spectrum estimates by a directional wave buoy.
{"title":"Ship Motion-Based Wave Estimation Using a Spectral Residual-Calculation","authors":"U. Nielsen, Astrid H. Brodtkorb","doi":"10.1109/oceanskobe.2018.8559136","DOIUrl":"https://doi.org/10.1109/oceanskobe.2018.8559136","url":null,"abstract":"This paper presents a study focused on a newly developed procedure for wave spectrum estimation using wave-induced motion recordings from a ship. The particular procedure stands out from other existing, similar ship motion-based procedures by its computational efficiency and - at the same time - providing accurate estimates of the on-site wave conditions. In the paper, the procedure is applied to full-scale experimental data obtained from dedicated sea trial runs. The results show favorable agreement with corresponding wave spectrum estimates by a directional wave buoy.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"92 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131032843","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8559276
Marielle Malfante, Omar Mohammed, C. Gervaise, M. Dalla Mura, J. Mars
The work presented in this paper focuses on the environmental monitoring of underwater areas using acoustic signals. In particular, we propose to compare the effectiveness of various feature sets used to represent the underwater acoustic data for the automatic processing of fish sounds We focus on the detection and classification tasks. Specifically, we compare the use of features issued from signal processing presented and validated in [15], [16] to the use of features obtained through deep convolutional neural networks. Experimental results show that the use of signal processing features outperform the deep features in terms of classification accuracy.
{"title":"Use of Deep Features for the Automatic Classification of Fish Sounds","authors":"Marielle Malfante, Omar Mohammed, C. Gervaise, M. Dalla Mura, J. Mars","doi":"10.1109/OCEANSKOBE.2018.8559276","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8559276","url":null,"abstract":"The work presented in this paper focuses on the environmental monitoring of underwater areas using acoustic signals. In particular, we propose to compare the effectiveness of various feature sets used to represent the underwater acoustic data for the automatic processing of fish sounds We focus on the detection and classification tasks. Specifically, we compare the use of features issued from signal processing presented and validated in [15], [16] to the use of features obtained through deep convolutional neural networks. Experimental results show that the use of signal processing features outperform the deep features in terms of classification accuracy.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116592443","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8558880
A. Martins, J. Almeida, C. Almeida, B. Matias, S. Kapusniak, E. Silva
This paper presents EVA, a new concept for an hybrid ROV/AUV designed to support the underwater operation of an underwater mining machine, developed in the context of the European H2020 R&D ¡VAMOS! Project. This project is briefly presented, introducing the main components and concepts, providing the reader with clear picture of the operational scenario and allowing to understand better the functionality requirements of the support robotic vehicle developed. The design of EVA is detailed presented, addressing the mechanical design, hardware architecture, sensor system and navigation and control. The results of EVA both in water test tank, in the !VAMOS! Field trials in Lee Moor, UK, and in an harbor scenario are presented and discussed
{"title":"EVA a Hybrid ROV/AUV for Underwater Mining Operations Support","authors":"A. Martins, J. Almeida, C. Almeida, B. Matias, S. Kapusniak, E. Silva","doi":"10.1109/OCEANSKOBE.2018.8558880","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8558880","url":null,"abstract":"This paper presents EVA, a new concept for an hybrid ROV/AUV designed to support the underwater operation of an underwater mining machine, developed in the context of the European H2020 R&D ¡VAMOS! Project. This project is briefly presented, introducing the main components and concepts, providing the reader with clear picture of the operational scenario and allowing to understand better the functionality requirements of the support robotic vehicle developed. The design of EVA is detailed presented, addressing the mechanical design, hardware architecture, sensor system and navigation and control. The results of EVA both in water test tank, in the !VAMOS! Field trials in Lee Moor, UK, and in an harbor scenario are presented and discussed","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"176 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121724701","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8558855
D. Curto, S. Neugebauer, A. Viola, M. Traverso, V. Franzitta, M. Trapanese
The work presents two different revolutionary devices for the utilization of a new entry of renewable energy sources: sea wave. The first technology is based on linear generators, able to directly converts a linear motion into electrical output, limiting to minimum the chain of energy conversion. The other solution is based on a mechanical motion converter, coupled with alternators. The scope of this paper is to compare the two different systems designed by University of Palermo, through Life Cycle Assessment, in order to evaluate the global effects of the two systems to the environment.
{"title":"First Life Cycle Impact Considerations of Two Wave Energy Converters","authors":"D. Curto, S. Neugebauer, A. Viola, M. Traverso, V. Franzitta, M. Trapanese","doi":"10.1109/OCEANSKOBE.2018.8558855","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8558855","url":null,"abstract":"The work presents two different revolutionary devices for the utilization of a new entry of renewable energy sources: sea wave. The first technology is based on linear generators, able to directly converts a linear motion into electrical output, limiting to minimum the chain of energy conversion. The other solution is based on a mechanical motion converter, coupled with alternators. The scope of this paper is to compare the two different systems designed by University of Palermo, through Life Cycle Assessment, in order to evaluate the global effects of the two systems to the environment.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123758812","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}
Plankton are critically important to our ecosystem, accounting for more than half the primary productivity on earth and nearly half the total carbon fixed in the global carbon cycle. Loss of plankton populations could result in ecological upheaval as well as negative societal impacts. By contrast, a bloom of phytoplankton can result in red tides which will cause huge economic loss. So it's a valuable thing for people to get the species population and distribution information. Recently, convolutional neural networks have achieved state of the art result on large scale image classification. We use several popular CNN models on WHOI large scale plankton database, it has achieved high accuracy on this dataset, but the data distribution of WHOI is not balance, so we have to solve a data imbalance problem. To evaluate the classier in an impartial way, we introduce an evaluation criterion called F1 score. Although the CNN method have achieved high global accuracy on the database, they achieved low F1 score: 0.17, 0.29 on CIFAR10 CNN model and VGG16 model separately. In this paper, we introduced a transfer parallel model approach to overcome this problem. We pre-trained a CNN model on the small classes which have images less than 5,000. Then the pre-trained model was treated as a feature extractor to enhance the small class's features and we fixed all the weights of this pre-trained model and combined with a parallel network to train on the whole training database. Through this transferred feature based approach we achieved high F1 score 0.3752, 0.5444 with our model based on CIFAR10 CNN model and VGG16 model respectively.
{"title":"Transferred Parallel Convolutional Neural Network for Large Imbalanced Plankton Database Classification","authors":"Chao Wang, Xueer Zheng, Chunfeng Guo, Zhibin Yu, Jia Yu, Haiyong Zheng, Bing Zheng","doi":"10.1109/OCEANSKOBE.2018.8558836","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8558836","url":null,"abstract":"Plankton are critically important to our ecosystem, accounting for more than half the primary productivity on earth and nearly half the total carbon fixed in the global carbon cycle. Loss of plankton populations could result in ecological upheaval as well as negative societal impacts. By contrast, a bloom of phytoplankton can result in red tides which will cause huge economic loss. So it's a valuable thing for people to get the species population and distribution information. Recently, convolutional neural networks have achieved state of the art result on large scale image classification. We use several popular CNN models on WHOI large scale plankton database, it has achieved high accuracy on this dataset, but the data distribution of WHOI is not balance, so we have to solve a data imbalance problem. To evaluate the classier in an impartial way, we introduce an evaluation criterion called F1 score. Although the CNN method have achieved high global accuracy on the database, they achieved low F1 score: 0.17, 0.29 on CIFAR10 CNN model and VGG16 model separately. In this paper, we introduced a transfer parallel model approach to overcome this problem. We pre-trained a CNN model on the small classes which have images less than 5,000. Then the pre-trained model was treated as a feature extractor to enhance the small class's features and we fixed all the weights of this pre-trained model and combined with a parallel network to train on the whole training database. Through this transferred feature based approach we achieved high F1 score 0.3752, 0.5444 with our model based on CIFAR10 CNN model and VGG16 model respectively.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125212192","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 : 2018-05-28DOI: 10.1109/OCEANSKOBE.2018.8559366
Cheng Chen, Kun-de Yang, Y. Liu
An abnormal multipath arrival structure was found in the signal from a receiver array during an acoustic experiment in East China Sea. The cause for the abnormal pattern was found due to the tilt of the array, which results from the strong sea currents. The array tilt was figured out with the Bellhop acoustic ray model by matching the simulation results with the received signal. Results were further testified by the depth sensor data. Given that the ocean observations are always sparse in time and spatial domain, the method in this paper provides a new source to help determine the sea current direction in the complex shallow water environment.
{"title":"Estimating the Array Tilt with the Received Signal in East China Sea","authors":"Cheng Chen, Kun-de Yang, Y. Liu","doi":"10.1109/OCEANSKOBE.2018.8559366","DOIUrl":"https://doi.org/10.1109/OCEANSKOBE.2018.8559366","url":null,"abstract":"An abnormal multipath arrival structure was found in the signal from a receiver array during an acoustic experiment in East China Sea. The cause for the abnormal pattern was found due to the tilt of the array, which results from the strong sea currents. The array tilt was figured out with the Bellhop acoustic ray model by matching the simulation results with the received signal. Results were further testified by the depth sensor data. Given that the ocean observations are always sparse in time and spatial domain, the method in this paper provides a new source to help determine the sea current direction in the complex shallow water environment.","PeriodicalId":441405,"journal":{"name":"2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129632777","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：