Automated Diatom Classification (Part B): A Deep Learning Approach

Pedraza, Aníbal; Bueno, Gloria; Deniz, Oscar; Cristóbal, Gabriel; Blanco, Saúl; Borrego-Ramos, María

doi:10.3390/app7050460

Cited by 104 publications

(106 citation statements)

References 11 publications

Supporting

Mentioning

100

Contrasting

Unclassified

Order By: Relevance

“…While data augmentation may introduce some bias in the experiments, mainly related to the invariant features, our aim in adding the augmentation was to compare classic methods with deep learning (see [24]), for which the same augmentation is done. Moreover, as mentioned above, most features (85%) are not invariant to rotation and mirroring, and therefore, data represented with these features can be considered without bias.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Automated Diatom Classification (Part A): Handcrafted Feature Approaches

et al. 2017

Self Cite

View full text Add to dashboard Cite

This paper deals with automatic taxa identification based on machine learning methods. The aim is therefore to automatically classify diatoms, in terms of pattern recognition terminology. Diatoms are a kind of algae microorganism with high biodiversity at the species level, which are useful for water quality assessment. The most relevant features for diatom description and classification have been selected using an extensive dataset of 80 taxa with a minimum of 100 samples/taxon augmented to 300 samples/taxon. In addition to published morphological, statistical and textural descriptors, a new textural descriptor, Local Binary Patterns (LBP), to characterize the diatom's valves, and a log Gabor implementation not tested before for this purpose are introduced in this paper. Results show an overall accuracy of 98.11% using bagging decision trees and combinations of descriptors. Finally, some phycological features of diatoms that are still difficult to integrate in computer systems are discussed for future work.

show abstract

Section: Resultsmentioning

confidence: 99%

“…They obtained an accuracy up to 96.3%. The other work related to CNN applied to diatoms is the one published by the authors (see the next paper companion [24]). The work presented here and the methodology have been compared to the CNN approach [24].…”

Section: Introductionmentioning

confidence: 99%

Automated Diatom Classification (Part A): Handcrafted Feature Approaches

et al. 2017

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the classical method part (Dorado 2016;Pedraza et al 2017a), shape feature extraction are first done, then SVM and RF classifiers are used to classify different diatom categories. In the novel method part (Dorado 2016;Pedraza et al 2017b), a deep learning framework (CNN) is designed using the Caffe and LeNet resources. In the experiment, 80 types of the WMs (100 samples/type) are used to test the effectiveness of the methods, and the best accuracy around 99.96% is obtained using the deep learning method.…”

Section: Original Methodsmentioning

confidence: 99%

A survey for the applications of content-based microscopic image analysis in microorganism classification domains

Wang

2017

Artif Intell Rev

View full text Add to dashboard Cite

Microorganisms such as protozoa and bacteria play very important roles in many practical domains, like agriculture, industry and medicine. To explore functions of different categories of microorganisms is a fundamental work in biological studies, which can assist biologists and related scientists to get to know more properties, habits and characteristics of these tiny but obbligato living beings. However, taxonomy of microorganisms (microorganism classification) is traditionally investigated through morphological, chemical or physical analysis, which is time and money consuming. In order to overcome this, since the 1970s innovative content-based microscopic image analysis (CBMIA) approaches are introduced to microbiological fields. CBMIA methods classify microorganisms into different categories using multiple artificial intelligence approaches, such as machine vision, pattern recognition and machine learning algorithms. Furthermore, because CBMIA approaches are semior full-automatic computer-based methods, they are very efficient and labour cost saving, supporting a technical feasibility for microorganism classification in our current big data age. In this article, we review the development history of microorganism classification using CBMIA approaches with two crossed pipelines. In the first pipeline, all related works are grouped by their corresponding microorganism application domains. By this pipeline, it is easy for microbiologists to have an insight into each special application domain and find their interested applied CBMIA techniques. In the second pipeline, the related works in each application domain are reviewed by time periods. Using this pipeline, computer scientists can see the dynamic of technological development clearly and keep up with the future devel-

show abstract

“…Existing instruments take images with a 10-20× magnification of cells. In recent studies, the differentiation of up to 80 species as maximum was demonstrated with machine learning (50). Since 2009, the ImageStream X Mk II (Luminex, Austin, Texas) IFC has been available and used mainly for biomedical research, only occasionally has it been used for phytoplankton analysis (45,46).…”

Section: Imaging Flow Cytometrymentioning

confidence: 99%

Hidden Secrets Behind Dots: Improved Phytoplankton Taxonomic Resolution Using High‐Throughput Imaging Flow Cytometry

Dunker

2019

Cytometry Pt A

View full text Add to dashboard Cite

Phytoplankton are aquatic, microscopically small primary producers, accounting for almost half of the worldwide carbon fixation. As early indicators of environmental change, they play a crucial role in water quality management. Human activities like climate change, eutrophication, or international shipping traffic strongly impact diversity of these organisms. Phytoplankton monitoring is a crucial step in the recognition of changes in community composition. The common standard for monitoring programs is manual microscopic counting, which strongly limits sample number and sampling frequency. In contrast, high-throughput technologies like standard flow cytometry (FCM) are restricted to a low taxonomic resolution, which makes them unsuitable for the identification of indicator species. Imaging flow cytometers (IFC) could overcome these limitations as they combine microscopy and high-throughput analysis. In comparison to single fluorescence values, image information not only allows for a wide variety of possibilities to characterize different species as well as immediate and fast measurements but also provides an archivable data output. Taxonomic resolution of IFC (ImageStream X Mk II) was proven comparable to standard FCM (FACSAria II) by the help of numerical evaluations. This is demonstrated on different levels of taxonomic differentiation of laboratory grown cultures in this study. Phytoplankton species discrimination by an imaging flow cytometer could be useful as supportive tool to make machine-learning classifications more robust, reliable, and flexible. Furthermore, this study provides examples, demonstrating the possibility of discrimination between species with similar fluorescence properties, strains, and even subpopulations. In contrast to standard FCM, each cell is not only represented as a dot in a cytogram but is also linked to microscopic brightfield and the author presents a new way to visualize this as image-based cytograms. The source code is supplied and could be useful for all kind of IFC data in general.

show abstract

Automated Diatom Classification (Part B): A Deep Learning Approach

Cited by 104 publications

References 11 publications

Automated Diatom Classification (Part A): Handcrafted Feature Approaches

Automated Diatom Classification (Part A): Handcrafted Feature Approaches

A survey for the applications of content-based microscopic image analysis in microorganism classification domains

Hidden Secrets Behind Dots: Improved Phytoplankton Taxonomic Resolution Using High‐Throughput Imaging Flow Cytometry

Contact Info

Product

Resources

About