PyTorchKR Papers

최신 논문

총 11,809편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1909.10567

Tangent space spatial filters for interpretable and efficient Riemannian classification

Methods based on Riemannian geometry have proven themselves to be good models for decoding in brain-computer interfacing (BCI). However, one major drawback of these methods is that it is not possible to determine what aspect of the signal the classifier is built on, leaving open the possibility that artifacts drive classification performance. In areas where artifactual control is problematic, specifically neurofeedback and BCIs in patient populations, this has led people to continue to rely on spatial filters as a way of generating features that are provably brain-related. Furthermore, these methods also suffer from the curse of dimensionality and are almost infeasible in high-density online BCI systems. To tackle these drawbacks, we introduce here a method for computing spatial filters from any linear function in the Riemannian tangent space, which allows for more efficient classification as well as the removal of artifact sources from classifiers built on Riemannian methods. We first prove a fundamental relationship between certain tangent spaces and spatial filtering methods, including an explanation of common spatial patterns within this framework, and then validate our proposed approach using an open-access BCI analysis framework.

arXiv:1909.06031

Deep Learning for Cooperative Radio Signal Classification

Radio signal classification has a very wide range of applications in cognitive radio networks and electromagnetic spectrum monitoring. In this article, we consider scenarios where multiple nodes in the network participate in cooperative classification. We propose cooperative radio signal classification methods based on deep learning for decision fusion, signal fusion and feature fusion, respectively. We analyze the performance of these methods through simulation experiments. We conclude the article with a discussion of research challenges and open problems.

arXiv:1909.05397

Multitask Classification and Segmentation for Cancer Diagnosis in Mammography

Annotation cost is a bottleneck for collecting massive data in mammography, especially for training deep neural networks. In this paper, we study the use of heterogeneous levels of annotation granularity to improve predictive performances. More precisely, we introduce a multi-task learning scheme for training convolutional neural network (ConvNets), which combines segmentation and classification, using image-level and pixel-level annotations. In this way, different objectives can be used to regularize training by sharing intermediate deep representations. Successful experiments are carried out on the Digital Database of Screening Mammography (DDSM) to validate the relevance of the proposed approach.

arXiv:1908.09401

Machine-learning enables Image Reconstruction and Classification in a "see-through" camera

We demonstrate that image reconstruction can be achieved via a convolutional neural network for a "see-through" computational camera comprised of a transparent window and a CMOS image sensor. Furthermore, we compared classification results using a classifier network for the raw sensor data vs the reconstructed images. The results suggest that similar classification accuracy is likely possible in both cases with appropriate network optimizations. All networks were trained and tested for the MNIST (6 classes), EMNIST and the Kanji49 datasets.

arXiv:1908.08476

WiFi Motion Detection: A Study into Efficacy and Classification

WiFi and security pose both an issue and act as a growing presence in everyday life. Today's motions detection implementations are severely lacking in the areas of secrecy, scope, and cost. To combat this problem, we aim to develop a motion detection system that utilizes WiFi Channel State Information (CSI), which describes how a wireless signal propagates from the transmitter to the receiver. The goal of this study is to develop a real-time motion detection and classification system that is discreet, cost-effective, and easily implementable. The system would only require an Ubuntu laptop with an Intel Ultimate N WiFi Link 5300 and a standard router. The system will be developed in two parts: (1) a robust system to track CSI variations in real-time, and (2) an algorithm to classify the motion. The system used to track CSI variance in real-time was completed in August 2018. Initial results show that introduction of motion to a previously motionless area is detected with high confidence. We present the development of (1) anomaly detection, utilizing the moving average filter implemented in the initial program and/or unsupervised machine learning, and (2) supervised machine learning algorithms to classify a set of simple motions using a proposed feature extraction methods. Lastly, classification methods such as Decision Tree, Naive Bayes, and Long Short-Term Memory can be used to classify basic actions regardless of speed, location, or orientation.

arXiv:1907.13409

Hierarchical Fine-Tuning for joint Liver Lesion Segmentation and Lesion Classification in CT

We present an automatic method for joint liver lesion segmentation and classification using a hierarchical fine-tuning framework. Our dataset is small, containing 332 2-D CT examinations with lesion annotated into 3 lesion types: cysts, hemangiomas, and metastases. Using a cascaded U-net that performs segmentation and classification simultaneously, we trained a strong lesion segmentation model on the dataset of MICCAI 2017 Liver Tumor Segmentation (LiTS) Challenge. We used the trained weights to fine-tune a slightly modified model to obtain improved lesion segmentation and classification, on the smaller dataset. Since pre-training was done with similar data on a related task, we were able to learn more representative features (especially higher-level features in the U-Net's encoder), and improve pixel-wise classification results. We show an improvement of over 10\% in Dice score and classification accuracy, compared to a baseline model. We further improve the classification performance by hierarchically freezing the encoder part of the network and achieve an improvement of over 15\% in Dice score and classification accuracy. We compare our results with an existing method and show an improvement of 14\% in the success rate and 12\% in the classification accuracy.

arXiv:1907.11953

Automated Mammogram Analysis with a Deep Learning Pipeline

Current deep learning based detection models tackle detection and segmentation tasks by casting them to pixel or patch-wise classification. To automate the initial mass lesion detection and segmentation on the whole mammographic images and avoid the computational redundancy of patch-based and sliding window approaches, the conditional generative adversarial network (cGAN) was used in this study. Subsequently, feeding the detected regions to the trained densely connected network (DenseNet), the binary classification of benign versus malignant was predicted. We used a combination of publicly available mammographic data repositories to train the pipeline, while evaluating the model's robustness toward our clinically collected repository, which was unseen to the pipeline.

arXiv:1906.09990

Self-repairing Classification Algorithms for Chemical Sensor Array

Chemical sensors are usually affected by drift, have low fabrication reproducibility and can experience failure or breaking events over the long term. Albeit improvements in fabrication processes are often slow and inadequate for completely surmounting these issues, data analysis can be used as of now to improve the available device performances. The present paper illustrates an algorithm, called Self-Repairing (SR), developed for repairing classification models after the occurrences of failures in sensor arrays. The procedure considers replacing broken sensors with replicas and eventually Self-Repairing algorithm trains these blank elements. Unlike the habitual alternatives reported in literature, SR performs this operation without the need of a whole new recalibration, references gas measurements or transfer dataset and, at the same time, without interrupting the on-going procedure of gas identification. Furthermore, Self-Repairing algorithm can utilize most of the standard classifiers as core algorithm; in this paper SR has been applied to k-NN, PLS-DA and LDA as examples. Models have been tested in a synthetic and real scenario considering sensor arrays affected by drift and eventually by failures. Real experiment has been performed with a set of metal oxide sensors over an 18-months period. Finally, the algorithm has been compared with standard version of chosen classifiers (k-NN, LDA and PLS-DA) showing superior performances of Self-Repairing and increasing the tolerance versus consecutive failures.

arXiv:1905.09059

Selection of a Minimal Number of Significant Porcine SNPs by an Information Gain and Genetic Algorithm Hybrid Model

A panel of large number of common Single Nucleotide Polymorphisms (SNPs) distributed across an entire porcine genome has been widely used to represent genetic variability of pig. With the advent of SNP-array technology, a genome-wide genetic profile of a specimen can be easily observed. Among the large number of such variations, there exist a much smaller subset of the SNP panel that could equally be used to correctly identify the corresponding breed. This work presents a SNP selection heuristic that can still be used effectively in the breed classification process. The proposed feature selection was done by the approach of combining a filter method and a wrapper method--information gain method and genetic algorithm--plus a feature frequency selection step, while classification was done by support vector machine. The approach was able to reduce the number of significant SNPs to 0.86 % of the total number of SNPs in a swine dataset and provided a high classification accuracy of 94.80 %.

arXiv:1905.00979공식 코드 · Papers with Code

City classification from multiple real-world sound scenes

The majority of sound scene analysis work focuses on one of two clearly defined tasks: acoustic scene classification or sound event detection. Whilst this separation of tasks is useful for problem definition, they inherently ignore some subtleties of the real-world, in particular how humans vary in how they describe a scene. Some will describe the weather and features within it, others will use a holistic descriptor like `park', and others still will use unique identifiers such as cities or names. In this paper, we undertake the task of automatic city classification to ask whether we can recognize a city from a set of sound scenes? In this problem each city has recordings from multiple scenes. We test a series of methods for this novel task and show that a simple convolutional neural network (CNN) can achieve accuracy of 50%. This is less than the acoustic scene classification task baseline in the DCASE 2018 ASC challenge on the same data. A simple adaptation to the class labels of pairing city labels with grouped scenes, accuracy increases to 52%, closer to the simpler scene classification task. Finally we also formulate the problem in a multi-task learning framework and achieve an accuracy of 56%, outperforming the aforementioned approaches.

arXiv:1904.00228

Voltage Quality Time Series Classification using Convolutional Neural Network

This paper presents the effectiveness of convolutional neural network (CNN) to classify power quality problems. These problems arise mainly due to increase in use of non-linear loads, operation of devices like adjustable speed drives and power factor correction capacitors, which is a growing concern both for utilities and customers. This work uses the advances in supervised learning to classify different power quality time-series waveforms such as voltage sag, swell, interruption, harmonics, transients and flicker. CNN results in a very high classification accuracy compared to other traditional and machine learning methods in presence of noise. This process can be employed by utilities as well as customers to understand the cause and mitigate voltage quality problems.

arXiv:1903.09606

Towards adversarial learning of speaker-invariant representation for speech emotion recognition

Speech emotion recognition (SER) has attracted great attention in recent years due to the high demand for emotionally intelligent speech interfaces. Deriving speaker-invariant representations for speech emotion recognition is crucial. In this paper, we propose to apply adversarial training to SER to learn speaker-invariant representations. Our model consists of three parts: a representation learning sub-network with time-delay neural network (TDNN) and LSTM with statistical pooling, an emotion classification network and a speaker classification network. Both the emotion and speaker classification network take the output of the representation learning network as input. Two training strategies are employed: one based on domain adversarial training (DAT) and the other one based on cross-gradient training (CGT). Besides the conventional data set, we also evaluate our proposed models on a much larger publicly available emotion data set with 250 speakers. Evaluation results show that on IEMOCAP, DAT and CGT provides 5.6% and 7.4% improvement respectively, over a baseline system without speaker-invariant representation learning on 5-fold cross validation. On the larger emotion data set, while CGT fails to yield better results than baseline, DAT can still provide 9.8% relative improvement on a standalone test set.

arXiv:1903.01907

Feature Selection for Transient Stability Assessment Based on Improved Maximal Relevance and Minimal Redundancy Criterion

A new feature selection method based on an improved maximal relevance and minimal redundancy (mRMR) criterion was proposed for power system transient stability assessment. First, the standard mRMR was improved by introducing a weight coefficient in the evaluation criteria to refine the measurement of the features correlation and redundancy. Then, the possible real-time information provided by phasor measurement unit (PMU) considered, a group of system-level classification features were extracted from the power system operation parameters to build the original feature set, and the improved mRMR was employed to evaluate the classification capability of the original features for feature selection. A group of nested candidate feature subsets were obtained by using the incremental search technique, and each candidate feature subset was tested by a support vector machine classifier to find the optimal feature subset with the highest classification accuracy. The effectiveness of the proposed method was validated by the simulation results on the New England 39-bus system and IEEE 50-generator test system.

arXiv:1812.04183

Completed Local Derivative Pattern for Rotation Invariant Texture Classification

In this paper, we propose a new texture descriptor, completed local derivative pattern (CLDP). In contrast to completed local binary pattern (CLBP), which involves only local differences at each scale, CLDP encodes the directional variation of the local differences of two scales as a complementary component to local patterns in CLBP. The new component in CLDP, with regarded as the directional derivative pattern, reflects the directional smoothness of local textures without increasing computation complexity. Experimental results on the Outex database show that CLDP, as a uni-scale pattern, outperforms uni-scale state-of-the-art texture descriptors on texture classification and has comparable performance with multi-scale texture descriptors.

arXiv:1812.04174

Scale Selective Extended Local Binary Pattern for Texture Classification

In this paper, we propose a new texture descriptor, scale selective extended local binary pattern (SSELBP), to characterize texture images with scale variations. We first utilize multi-scale extended local binary patterns (ELBP) with rotation-invariant and uniform mappings to capture robust local micro- and macro-features. Then, we build a scale space using Gaussian filters and calculate the histogram of multi-scale ELBPs for the image at each scale. Finally, we select the maximum values from the corresponding bins of multi-scale ELBP histograms at different scales as scale-invariant features. A comprehensive evaluation on public texture databases (KTH-TIPS and UMD) shows that the proposed SSELBP has high accuracy comparable to state-of-the-art texture descriptors on gray-scale-, rotation-, and scale-invariant texture classification but uses only one-third of the feature dimension.

arXiv:1812.00888

A Consolidated Approach to Convolutional Neural Networks and the Kolmogorov Complexity

The ability to precisely quantify similarity between various entities has been a fundamental complication in various problem spaces specifically in the classification of cellular images. Contemporary similarity measures applied in the domain of image processing proposed by the scientific community are mainly pursued in supervised settings. In this work, we will explore the innovative algorithmic normalized compression distance metric based on the information theoretic concept of Kolmogorov Complexity. Additionally we will observe its possible implementation in Convolutional Neural Networks to facilitate and automate the classification of Retinal Pigment Epithelial cell cultures for use in Age Related Macular Degeneration Stem Cell therapy in an unsupervised setting.

arXiv:1810.10329

Vehicle classification using ResNets, localisation and spatially-weighted pooling

We investigate whether ResNet architectures can outperform more traditional Convolutional Neural Networks on the task of fine-grained vehicle classification. We train and test ResNet-18, ResNet-34 and ResNet-50 on the Comprehensive Cars dataset without pre-training on other datasets. We then modify the networks to use Spatially Weighted Pooling. Finally, we add a localisation step before the classification process, using a network based on ResNet-50. We find that using Spatially Weighted Pooling and localisation both improve classification accuracy of ResNet50. Spatially Weighted Pooling increases accuracy by 1.5 percent points and localisation increases accuracy by 3.4 percent points. Using both increases accuracy by 3.7 percent points giving a top-1 accuracy of 96.351\% on the Comprehensive Cars dataset. Our method achieves higher accuracy than a range of methods including those that use traditional CNNs. However, our method does not perform quite as well as pre-trained networks that use Spatially Weighted Pooling.

arXiv:1810.10062

Recognition of basic hand movements using Electromyography

The aim of this work was to identify six basic movements of the hand using two systems. Being an interdisciplinary topic, there has been conducted studying in the anatomy of forearm muscles, biosignals, the method of electromyography (EMG) and methods of pattern recognition. Moreover, the signal contained enough noise and had to be analyzed, using EMD, to extract features and to reduce its dimensionality, using RELIEF and PCA, to improve the success rate of classification. The first part uses an EMG system of Delsys initially for an individual and then for six people with the average successful classification, for these six movements at rates of over 80%. The second part involves the construction of an autonomous system EMG using an Arduino microcontroller, EMG sensors and electrodes, which are arranged in an elastic glove. Classification results in this case reached 75% of success.

arXiv:1810.04121

Inter-Patient ECG Classification with Convolutional and Recurrent Neural Networks

The recent advances in ECG sensor devices provide opportunities for user self-managed auto-diagnosis and monitoring services over the internet. This imposes the requirements for generic ECG classification methods that are inter-patient and device independent. In this paper, we present our work on using the densely connected convolutional neural network (DenseNet) and gated recurrent unit network (GRU) for addressing the inter-patient ECG classification problem. A deep learning model architecture is proposed and is evaluated using the MIT-BIH Arrhythmia and Supraventricular Databases. The results obtained show that without applying any complicated data pre-processing or feature engineering methods, both of our models have considerably outperformed the state-of-the-art performance for supraventricular (SVEB) and ventricular (VEB) arrhythmia classifications on the unseen testing dataset (with the F1 score improved from 51.08 to 61.25 for SVEB detection and from 88.59 to 89.75 for VEB detection respectively). As no patient-specific or device-specific information is used at the training stage in this work, it can be considered as a more generic approach for dealing with scenarios in which varieties of ECG signals are collected from different patients using different types of sensor devices.

arXiv:1810.00681

Learning Robust, Transferable Sentence Representations for Text Classification

Despite deep recurrent neural networks (RNNs) demonstrate strong performance in text classification, training RNN models are often expensive and requires an extensive collection of annotated data which may not be available. To overcome the data limitation issue, existing approaches leverage either pre-trained word embedding or sentence representation to lift the burden of training RNNs from scratch. In this paper, we show that jointly learning sentence representations from multiple text classification tasks and combining them with pre-trained word-level and sentence level encoders result in robust sentence representations that are useful for transfer learning. Extensive experiments and analyses using a wide range of transfer and linguistic tasks endorse the effectiveness of our approach.