A review of deep learning in image recognition

Pak, Myeongsuk; Kim, Sang-Hoon

doi:10.1109/caipt.2017.8320684

Cited by 157 publications

(69 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Indeed, supervised machine learning using deep Convolutional Neural Networks (CNNs) and, in particular, deep Residual neural Networks (ResNet) have been shown to perform optimally for the classification of images (6,10) . To leverage the power of ResNet classifiers, we converted the raw signal corresponding to the extracted barcodes into an array of pixels using diverse image transformation strategies previously shown to be effective for subsequent CNN training and classification, including recurrence plots (RP) (11) , Markov Transition Fields (MTF), Gramian Angular Difference Fields (GADF) and Gramian Angular Summation Fields (GASF) (12) .…”

Section: Transformation Of Segmented Barcode Signals Into 2d Imagesmentioning

confidence: 99%

Barcoding and demultiplexing Oxford Nanopore native RNA sequencing reads with deep residual learning

Smith

Ersavas

Ferguson

et al. 2019

Preprint

View full text Add to dashboard Cite

Nanopore sequencing has enabled sequencing of native RNA molecules without conversion to cDNA, thus opening the gates to a new era for the unbiased study of RNA biology. However, a formal barcoding protocol for direct sequencing of native RNA molecules is currently lacking, limiting the efficient processing of multiple samples in the same flowcell. A major limitation for the development of barcoding protocols for direct RNA sequencing is the error rate introduced during the base-calling process, especially towards the 5' and 3' ends of reads, which complicates sequence-based barcode demultiplexing. Here, we propose a novel strategy to barcode and demultiplex direct RNA sequencing nanopore data, which does not rely on base-calling or additional library preparation steps.Specifically, custom DNA oligonucleotides are ligated to RNA transcripts during library preparation.Then, raw current signal corresponding to the DNA barcode is extracted and transformed into an array of pixels, which is used to determine the underlying barcode using a deep convolutional neural network classifier. Our method, DeePlexiCon , implements a 20-layer residual neural network model that can demultiplex 93% of the reads with 95.1% specificity, or 60% of reads with 99.9% specificity.The availability of an efficient and simple barcoding strategy for native RNA sequencing will enhance the use of direct RNA sequencing by making it more cost-effective to the entire community. Moreover, it will facilitate the applicability of direct RNA sequencing to samples where the RNA amounts are limited, such as patient-derived samples.

show abstract

Section: Transformation Of Segmented Barcode Signals Into 2d Imagesmentioning

confidence: 99%

Barcoding and demultiplexing Oxford Nanopore native RNA sequencing reads with deep residual learning

Smith

Ersavas

Ferguson

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

“…Various models and learning methods have been developed. The depth of deep learning networks has expanded from tens to hundreds [12], in contrast to conventional neural networks with a depth of two to three. Deep learning networks abstract data to a high level through a combination of nonlinear transforms.…”

Section: Introductionmentioning

confidence: 99%

“…Since posture recognition is performed using images, it can be applied to inexpensive cameras, and the device used for acquiring experimental data also has an inexpensive feature even though it supports a depth camera. In conventional machine learning, there is a limitation to recognizing posture directly using only an image [12,[35][36][37]. However, owing to advancements in deep learning, good performance in posture recognition can be achieved using only one image.…”

Section: Introductionmentioning

confidence: 99%

Posture Recognition Using Ensemble Deep Models under Various Home Environments

et al. 2020

View full text Add to dashboard Cite

This paper is concerned with posture recognition using ensemble convolutional neural networks (CNNs) in home environments. With the increasing number of elderly people living alone at home, posture recognition is very important for helping elderly people cope with sudden danger. Traditionally, to recognize posture, it was necessary to obtain the coordinates of the body points, depth, frame information of video, and so on. In conventional machine learning, there is a limitation in recognizing posture directly using only an image. However, with advancements in the latest deep learning, it is possible to achieve good performance in posture recognition using only an image. Thus, we performed experiments based on VGGNet, ResNet, DenseNet, InceptionResNet, and Xception as pre-trained CNNs using five types of preprocessing. On the basis of these deep learning methods, we finally present the ensemble deep model combined by majority and average methods. The experiments were performed by a posture database constructed at the Electronics and Telecommunications Research Institute (ETRI), Korea. This database consists of 51,000 images with 10 postures from 51 home environments. The experimental results reveal that the ensemble system by InceptionResNetV2s with five types of preprocessing shows good performance in comparison to other combination methods and the pre-trained CNN itself. 15 of 26 preprocessed posture images for the testing was 500 for standing normal, 497 for standing bent, 500 for sitting sofa, 500 for sitting chair, 497 for sitting floor, 498 for sitting squat, 489 for lying face, 489 for lying back, 482 for lying side, 481 for lying crouched, and 2001 for other images. The total number of images for the training, validation, and testing was 44,874, 7913, and 6934, respectively. detector [39], and the segmented person images were filtered manually. The number of preprocessed posture images for the training was 3910 for standing normal, 3802 for standing bent, 3899 for sitting sofa, 3899 for sitting chair, 3896 for sitting floor, 3859 for sitting squat, 3752 for lying face, 3777 for lying back, 3773 for lying side, 3505 for lying crouched, and 6802 for other images. The number of preprocessed posture images for the validation was 689 for standing normal, 670 for standing bent, 688 for sitting sofa, 687 for sitting chair, 687 for sitting floor, 681 for sitting squat, 662 for lying face, 666 for lying back, 665 for lying side, 618 for lying crouched, and 1200 for other images. The number of preprocessed posture images for the testing was 500 for standing normal, 497 for standing bent, 500 for sitting sofa, 500 for sitting chair, 497 for sitting floor, 498 for sitting squat, 489 for lying face, 489 for lying back, 482 for lying side, 481 for lying crouched, and 2001 for other images. The total number of images for the training, validation, and testing was 44,874, 7913, and 6934, respectively.

show abstract

“…The availability of a large number of instances of the protein space (be that sequence or structure) and the * raffaello.potestio@unitn.it necessity to perform dataset-wide analyses and screening of their properties naturally leads one to wonder whether one could take advantage of the recent progresses achieved by machine learning approaches, in particular deep learning (DL). The latter is a subset of the wide class of machine learning computational methods, and has been successfully applied to a fairly wide spectrum of areas of science [9,10], ranging from neuroscience [11] to image and speech recognition [12,13]. In the field of Computational Chemistry much effort has been devoted to the identification of the variables that are able to provide a comprehensive description of a chemical compound (molecular descriptors).…”

Section: Introductionmentioning

confidence: 99%

A deep learning approach to the structural analysis of proteins

Giulini

Potestio

2019

Interface Focus.

View full text Add to dashboard Cite

Deep learning (DL) algorithms hold great promise for applications in the field of computational biophysics. In fact, the vast amount of available molecular structures, as well as their notable complexity, constitutes an ideal context in which DL-based approaches can be profitably employed. To express the full potential of these techniques, though, it is a prerequisite to express the information contained in a molecule’s atomic positions and distances in a set of input quantities that the network can process. Many of the molecular descriptors devised so far are effective and manageable for relatively small structures, but become complex and cumbersome for larger ones. Furthermore, most of them are defined locally, a feature that could represent a limit for those applications where global properties are of interest. Here, we build a DL architecture capable of predicting non-trivial and intrinsically global quantities, that is, the eigenvalues of a protein’s lowest-energy fluctuation modes. This application represents a first, relatively simple test bed for the development of a neural network approach to the quantitative analysis of protein structures, and demonstrates unexpected use in the identification of mechanically relevant regions of the molecule.

show abstract

A review of deep learning in image recognition

Cited by 157 publications

References 6 publications

Barcoding and demultiplexing Oxford Nanopore native RNA sequencing reads with deep residual learning

Barcoding and demultiplexing Oxford Nanopore native RNA sequencing reads with deep residual learning

Posture Recognition Using Ensemble Deep Models under Various Home Environments

A deep learning approach to the structural analysis of proteins

Contact Info

Product

Resources

About