Dataset Information

An acquisition, curation and management workflow for sustainable, terabyte-scale marine image analysis.

ABSTRACT: Optical imaging is a common technique in ocean research. Diving robots, towed cameras, drop-cameras and TV-guided sampling gear: all produce image data of the underwater environment. Technological advances like 4K cameras, autonomous robots, high-capacity batteries and LED lighting now allow systematic optical monitoring at large spatial scale and shorter time but with increased data volume and velocity. Volume and velocity are further increased by growing fleets and emerging swarms of autonomous vehicles creating big data sets in parallel. This generates a need for automated data processing to harvest maximum information. Systematic data analysis benefits from calibrated, geo-referenced data with clear metadata description, particularly for machine vision and machine learning. Hence, the expensive data acquisition must be documented, data should be curated as soon as possible, backed up and made publicly available. Here, we present a workflow towards sustainable marine image analysis. We describe guidelines for data acquisition, curation and management and apply it to the use case of a multi-terabyte deep-sea data set acquired by an autonomous underwater vehicle.

SUBMITTER: Schoening T

PROVIDER: S-EPMC6111891 | biostudies-literature | 2018 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

An acquisition, curation and management workflow for sustainable, terabyte-scale marine image analysis.

Schoening Timm T Köser Kevin K Greinert Jens J

Scientific data 20180828

Optical imaging is a common technique in ocean research. Diving robots, towed cameras, drop-cameras and TV-guided sampling gear: all produce image data of the underwater environment. Technological advances like 4K cameras, autonomous robots, high-capacity batteries and LED lighting now allow systematic optical monitoring at large spatial scale and shorter time but with increased data volume and velocity. Volume and velocity are further increased by growing fleets and emerging swarms of autonomou ...[more]

PMID: 30152813

Similar Datasets

Project description:PurposeThe curation of images using human resources is time intensive but an essential step for developing artificial intelligence (AI) algorithms. Our goal was to develop and implement an AI algorithm for image curation in a high-volume setting. We also explored AI tools that will assist in deploying a tiered approach, in which the AI model labels images and flags potential mislabels for human review.DesignImplementation of an AI algorithm.ParticipantsSeven-field stereoscopic images from multiple clinical trials.MethodsThe 7-field stereoscopic image protocol includes 7 pairs of images from various parts of the central retina along with images of the anterior part of the eye. All images were labeled for field number by reading center graders. The model output included classification of the retinal images into 8 field numbers. Probability scores (0-1) were generated to identify misclassified images, with 1 indicating a high probability of a correct label.Main outcome measuresAgreement of AI prediction with grader classification of field number and the use of probability scores to identify mislabeled images.ResultsThe AI model was trained and validated on 17 529 images and tested on 3004 images. The pooled agreement of field numbers between grader classification and the AI model was 88.3% (kappa, 0.87). The pooled mean probability score was 0.97 (standard deviation [SD], 0.08) for images for which the graders agreed with the AI-generated labels and 0.77 (SD, 0.19) for images for which the graders disagreed with the AI-generated labels (P < 0.0001). Using receiver operating characteristic curves, a probability score of 0.99 was identified as a cutoff for distinguishing mislabeled images. A tiered workflow using a probability score of < 0.99 as a cutoff would include 27.6% of the 3004 images for human review and reduce the error rate from 11.7% to 1.5%.ConclusionsThe implementation of AI algorithms requires measures in addition to model validation. Tools to flag potential errors in the labels generated by AI models will reduce inaccuracies, increase trust in the system, and provide data for continuous model development.

Project description:Underwater video monitoring systems are being widely used in fisheries to investigate fish behavior in relation to fishing gear and fishing gear performance during fishing. Such systems can be useful to evaluate the catch composition as well. In demersal trawl fisheries, however, their applicability can be challenged by low light conditions, mobilized sediment and scattering in murky waters. In this study, we introduce a novel observation system (called NepCon) which aims at reducing current limitations by combining an optimized image acquisition setup and tailored image analyses software. The NepCon system includes a high-contrast background to enhance the visibility of the target objects, a compact camera and an artificial light source. The image analysis software includes a machine learning algorithm which is evaluated here to test automatic detection and count of Norway lobster (Nephrops norvegicus). NepCon is specifically designed for applications in demersal trawls and this first phase aims at increasing the accuracy of N. norvegicus detection at the data acquisition level. To find the best contrasting background for the purpose we compared the output of four image segmentation methods applied to static images of N. norvegicus fixed in front of four test background colors. The background color with the best performance was then used to evaluate computer vision and deep learning approaches for automatic detection, tracking and counting of N. norvegicus in the videos. In this initial phase we tested the system in an experimental setting to understand the feasibility of the system for future implementation in real demersal fishing conditions. The N. norvegicus directed trawl fishery typically has no assistance from underwater observation technology and therefore are largely conducted blindly. The demonstrated perception system achieves 76% accuracy (F-score) in automatic detection and count of N. norvegicus, which provides a significant elevation of the current benchmark.

Project description:Digitisation of natural history collections has evolved from creating databases for the recording of specimens' catalogue and label data to include digital images of specimens. This has been driven by several important factors, such as a need to increase global accessibility to specimens and to preserve the original specimens by limiting their manual handling. The size of the collections pointed to the need of high throughput digitisation workflows. However, digital imaging of large numbers of fragile specimens is an expensive and time-consuming process that should be performed only once. To achieve this, the digital images produced need to be useful for the largest set of applications possible and have a potentially unlimited shelf life. The constraints on digitisation speed need to be balanced against the applicability and longevity of the images, which, in turn, depend directly on the quality of those images. As a result, the quality criteria that specimen images need to fulfil influence the design, implementation and execution of digitisation workflows. Different standards and guidelines for producing quality research images from specimens have been proposed; however, their actual adaptation to suit the needs of different types of specimens requires further analysis. This paper presents the digitisation workflow implemented by Meise Botanic Garden (MBG). This workflow is relevant because of its modular design, its strong focus on image quality assessment, its flexibility that allows combining in-house and outsourced digitisation, processing, preservation and publishing facilities and its capacity to evolve for integrating alternative components from different sources. The design and operation of the digitisation workflow is provided to showcase how it was derived, with particular attention to the built-in audit trail within the workflow, which ensures the scalable production of high-quality specimen images and how this audit trail ensures that new modules do not affect either the speed of imaging or the quality of the images produced.

Project description:BackgroundThe development of high-throughput experimental technologies, such as next-generation sequencing, have led to new challenges for handling, analyzing and integrating the resulting large and diverse datasets. Bioinformatical analysis of these data commonly requires a number of mutually dependent steps applied to numerous samples for multiple conditions and replicates. To support these analyses, a number of workflow management systems (WMSs) have been developed to allow automated execution of corresponding analysis workflows. Major advantages of WMSs are the easy reproducibility of results as well as the reusability of workflows or their components.ResultsIn this article, we present Watchdog, a WMS for the automated analysis of large-scale experimental data. Main features include straightforward processing of replicate data, support for distributed computer systems, customizable error detection and manual intervention into workflow execution. Watchdog is implemented in Java and thus platform-independent and allows easy sharing of workflows and corresponding program modules. It provides a graphical user interface (GUI) for workflow construction using pre-defined modules as well as a helper script for creating new module definitions. Execution of workflows is possible using either the GUI or a command-line interface and a web-interface is provided for monitoring the execution status and intervening in case of errors. To illustrate its potentials on a real-life example, a comprehensive workflow and modules for the analysis of RNA-seq experiments were implemented and are provided with the software in addition to simple test examples.ConclusionsWatchdog is a powerful and flexible WMS for the analysis of large-scale high-throughput experiments. We believe it will greatly benefit both users with and without programming skills who want to develop and apply bioinformatical workflows with reasonable overhead. The software, example workflows and a comprehensive documentation are freely available at www.bio.ifi.lmu.de/watchdog.

Dataset Information

An acquisition, curation and management workflow for sustainable, terabyte-scale marine image analysis.

Publications

An acquisition, curation and management workflow for sustainable, terabyte-scale marine image analysis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets