Automated Emotion Classification in Free-moving Rats

2021 ◽  
Author(s):  
Andre Telfer
Keyword(s):  
Short Term Memory ◽  
Body Part ◽  
Emotion Classification ◽  
Recent Success ◽  
Emotion Label ◽  
Varying Environments ◽  
Artificial Neural Network Ann ◽  
Ann Classifier ◽  
Observer Drift ◽  
Future Work

Studies involving emotion often use animal models and currently rely on manual labelling by researchers. This human-driven labelling approach leads to a number of challenges such as: long analysis times, imprecise results, observer drift, and varying correlation between observers. These problems impact reproducibility, and have contributed to our lack of understanding of fundamental mechanical questions such as how emotions arise from neuronal circuits. Recent success of machine learning models across similar problems show that it can help to mitigate these challenges while meeting or exceeding human accuracy.  We developed a classifier pipeline that takes in videos and produces an emotion label. The pipeline extracts body part positions from each frame using a pose estimator and feeds them into an Artificial Neural Network (ANN) classifier built using stacked Long Short Term Memory (LSTM) layers. The data was collected by treating nine rats with Lypopolysaccharide (LPS) injections (10mg/kg). First, rats were recorded for 10 minutes under control conditions with no manipulation and no observed symptoms of stress or malaise. A week later, rats were injected with LPS and filmed for 10 minutes two hours post-injection.  The classifier pipeline developed correctly labelled 78% of the 125,040 video segments from 8 test videos. When combined with a vote-based system, this led to 7 of the 8 test videos being classified correctly which was the same accuracy attained by a human expert from the lab. The test videos had varying environments and used rats that were different from the training videos, providing evidence of a degree of robustness in the model. Future work will focus on expanding the test data and incorporating models for 3D pose estimation and behavioral classification.

A New Islanding Detection Method Based on Wavelet-transform and ANN for Micro-grid Including Inverter Assisted Distributed Generator

2019 ◽  
Vol 20 (5) ◽  
Author(s):  
Zhengyuan Guan ◽  
Yuan Liao
Keyword(s):  
Wavelet Transform ◽  
Harmonic Distortion ◽  
Distribution Networks ◽  
Islanding Detection ◽  
Detection Zone ◽  
Distributed Generator ◽  
Voltage Unbalance ◽  
Typical Distribution ◽  
Artificial Neural Network Ann ◽  
Ann Classifier

Abstract This paper presents a new composite approach based on wavelet-transform and ANN for islanding detection of distributed generation (DG). The proposed method first uses wavelet-transform to detect the occurrence of events, and then uses artificial neural network (ANN) to classify islanding and non-islanding events. Total harmonic distortion and voltage unbalance are extracted as feature inputs for ANN classifier. The performance of the proposed method is tested by simulations for two typical distribution networks based on MATLAB/Simulink. The results show that the developed method can effectively detect islanding with low misclassification. The method has the advantages of small non-detection zone and robustness against noises.

scholarly journals ANN-Based Soft Sensor to Predict Effluent Violations in Wastewater Treatment Plants

Sensors ◽  
2019 ◽  
Vol 19 (6) ◽  
pp. 1280 ◽  
Author(s):  
Ivan Pisa ◽  
Ignacio Santín ◽  
Jose Vicario ◽  
Antoni Morell ◽  
Ramon Vilanova
Keyword(s):  
Wastewater Treatment ◽  
Short Term Memory ◽  
Wastewater Treatment Plants ◽  
Control Strategies ◽  
False Positive Rate ◽  
Soft Sensor ◽  
Aquatic Life ◽  
Pollutant Concentrations ◽  
High Concentrations ◽  
Artificial Neural Network Ann

Wastewater treatment plants (WWTPs) form an industry whose main goal is to reduce water’s pollutant products, which are harmful to the environment at high concentrations. In addition, regulations are applied by administrations to limit pollutant concentrations in effluent. In this context, control strategies have been adopted by WWTPs to avoid violating these limits; however, some violations still occur. For that reason, this work proposes the deployment of an artificial neural network (ANN)-based soft sensor in which a Long-Short Term Memory (LSTM) network is used to generate predictions of nitrogen-derived components, specifically ammonium ( S N H ) and total nitrogen ( S N t o t ). S N t o t is a limiting nutrient and can therefore cause eutrophication, while nitrogen in the S N H form is toxic to aquatic life. These parameters are used by control strategies to allow actions to be taken in advance and only when violations are predicted. Since predictions complement control strategies, the evaluation of the ANN-based soft sensor was carried out using the Benchmark Simulation Model N.2. (BSM2) and three different control strategies (from low to high control complexity). Results show that our proposed method is able to predict nitrogen-derived products with good accuracy: the probability of detecting violations of BSM2’s limits is 86%–94%. Moreover, the prediction accuracy can be improved by calibrating the soft sensor; for example, perfect prediction of all future violations can be achieved at the expense of increasing the false positive rate.

Brazilian Coins Recognition Using Histogram of Oriented Gradients Features

2019 ◽  
Vol 16 (10) ◽  
pp. 4170-4178
Author(s):  
Sheifali Gupta ◽  
Gurleen Kaur ◽  
Deepali Gupta ◽  
Udit Jindal
Keyword(s):  
Region Of Interest ◽  
Initial Step ◽  
Support Vector ◽  
Histogram Of Oriented Gradients ◽  
Lighting Conditions ◽  
Roi Extraction ◽  
Feature Values ◽  
Artificial Neural Network Ann ◽  
Ann Classifier

This paper tends to the issue of coin recognition when dealing with shading and reflection variations under the same lighting conditions. In order to approach the problem, a database containing Brazilian coin images (both front and reverse side of the coin) consisting of five different denominations have been used which is provided by the kaggle-diverse and largest data community in the world. This work focuses on an automatic image classification process for Brazilian coins. The imagebased classification of coins primarily incorporates three stages where the initial step is Region of Interest (ROI) extraction; the subsequent advance is extraction of features and classification. The first step of ROI extraction is accomplished by segmenting the coin region using the proposed segmentation method. In the second step i.e., feature extraction; Histogram of Oriented Gradients (HOG) features are extracted from the image. The image is converted to a vector containing feature values. The third step is where the extracted features are mapped to the class and are known as classification. Three classification algorithms i.e., Support Vector Machine (SVM), Artificial Neural Network (ANN) and K-Nearest Neighbour are compared for classification of five coin denominations. With the proposed segmentation methodology, the best classification accuracy of 92% is achieved in the case of ANN classifier.

A COMBINED APPROACH TO TEXT-DEPENDENT SPEAKER IDENTIFICATION: COMPARISON WITH PURE NEURAL NET APPROACHES

1998 ◽  
Vol 08 (02) ◽  
pp. 273-281
Author(s):  
QIANHUI LIANG ◽  
MIAOLIANG ZHU
Keyword(s):  
Dynamic Time Warping ◽  
Speaker Identification ◽  
Neural Net ◽  
Combined Approach ◽  
Time Warping ◽  
Novel Approach ◽  
Artificial Neural Network Ann ◽  
Ann Classifier ◽  
Dynamic Time ◽  
Combined Classifier

A novel approach to automatic speaker identification (ASI) is presented. Most of the present automatic speaker identification systems based on neural networks have no definite mechanisms to compensate for time distortions due to elocution. Such models have less precise information about the intraspeaker measure. The new combined approach uses both distortion-based and discriminant-based methods. The distortion-based and discriminant-based methods are dynamic time warping (DTW) and artificial neural network (ANN) respectively. This paper compares this new classifier with a pure neural net classifier for speaker identification. The performance of the combined classifier surpasses that of a pure ANN classifier for the conditions tested.

scholarly journals Comparative Study of Wet and Dry Systems using EEG-based Convolution Neural Network

2020 ◽  
Author(s):  
Taweesak Emsawas ◽  
Tsukasa Kimura ◽  
Ken-ichi Fukui ◽  
Masayuki Numao
Keyword(s):  
Neural Network ◽  
Comparative Study ◽  
Short Term Memory ◽  
Model Performance ◽  
Ease Of Use ◽  
Convolution Neural Network ◽  
Cognitive State ◽  
Emotional States ◽  
Communication Tool ◽  
Emotion Classification

Abstract Brain-Computer Interface (BCI) is a communication tool between humans and systems using electroencephalography (EEG) to predict certain cognitive state aspects, such as attention or emotion. For brainwave recording, there are many types of acquisition devices created for different purposes. The wet system conducts the recording with electrode gel and can obtain high-quality brainwave signals, while the dry system expressly proposes the practical and ease of use. In this paper, we study a comparative study of wet and dry systems using two cognitive tasks: attention and music-emotion. The 3-back task is used as an assessment to measure attention and working memory in attention studies. Comparatively, the music-emotion experiments are used to predict the emotion according to the subject's questionnaires. Our analysis shows the similarities and differences between dry and wet electrodes by calculating the statistical values and frequency bands. Besides, we further study the relative characteristics by conducting the classification experiments. We proposed the end-to-end models of EEG classification, which are constructed by combining EEG-based feature extractors and classification networks. A deep convolution neural network (Deep ConvNet) and a shallow convolution neural network (Shallow ConvNet) were applied as the feature extractor of temporal and spatial filtering from raw EEG signals. The extracted feature is then forwardly conveyed to a long short-term memory ( LSTM ) to learn the dependencies of convolved features and classify attention states or emotional states. Additionally, transfer learning was utilized to improve the performance of the dry system by using transferred knowledge from the wet system. We applied the model not only on our dataset but also on the existing dataset to verify the model performance compared with the baseline techniques and the-state-of-the-art models. Using our proposed model, the result shows the significant differences between accuracy and chance level in attention classification (92.0%, S.D. 6.8%) and SEED dataset's emotion classification (75.3%, S.D. 9.3%).

scholarly journals Artifact Removal and of EEG Signal Classification for Brain Computer Interface (BCI) using Back Propagation

2020 ◽  
Vol 8 (6) ◽  
pp. 1275-1282
Keyword(s):  
Brain Computer Interface ◽  
Research Work ◽  
Back Propagation ◽  
Software Tool ◽  
Computer Interface ◽  
Eeg Signals ◽  
Biological Sources ◽  
Artificial Neural Network Ann ◽  
Ann Classifier ◽  
The Brain

A brain-computer interface (BCI) provides a communication passage between the brain and an external stratagem. The Brain and its EEG signals are acquired from the BCI along its control signals and its widely used mechanism in the field of the biomedical fields. In this research work, an artifacts are removed algorithm in the EEG is developed and simulated in the MATLAB 2017a software tool. EEG signals from patients are recoded while recording some of the artificial signals added to it, which are instigated by using eye blinks, eye movement, muscle, and cardiac noise, and also non-biological sources. Using suitable filters these artificial signals can be removed. This paper aims to remove the artificial signals from EEG signals and parameters like mean, standard. Deviation are calculated and compared with other methods such as LAMICA and FASTERs. In the paper, it is also the proposed arrangement of EEG signals for the discovery of typical and anomalous exercises utilizing Wavelet change and Artificial Neural Network (ANN) Classifier is considered. Here, the framework utilizes the back proliferation with feed-forward for order which pursues the ANN grouping. Accuracy of the classification is calculated and compared with other states of art publications and found that it is better.

scholarly journals Fast Detection and Traceback-based Mitigation of Interest Flooding Attack

2021 ◽  
Author(s):  
Naveen Kumar ◽  
Shashank Srivast
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Learning Approaches ◽  
Previous Approach ◽  
State Information ◽  
Fast Detection ◽  
Flooding Attack ◽  
Artificial Neural Network Ann ◽  
Ann Classifier ◽  
Better Than

Abstract NDN Pending Interest Table (PIT) helps NDN by storing the state of a request within the router. This state information helps the router to redirect the data packet towards the requester. However, an attacker can send malicious requests, which could flood the PIT; this attack is known as Interest Flooding Attack (IFA). In our previous work, we have found the most relevant features needed to detect IFA and applied a few machine learning approaches for the offline detection of IFA. In this article, a trained Artificial Neural Network (ANN) classifier has been deployed on each NDN router for the online detection of IFA. A novel traceback-based mitigation is proposed, which is triggered after the detection. The proposed approach is found better than the previous approach in terms of the satisfaction ratio and throughput of the legitimate consumers.

scholarly journals Mammogram Classification using Curvelet Coefficients and Gray Level Co-Occurrence Matrix for Detection of Breast Cancer

2019 ◽  
Vol 8 (12) ◽  
pp. 4819-4824
Keyword(s):  
Breast Cancer ◽  
Curvelet Transform ◽  
Test Methods ◽  
Computer Aided Diagnosis ◽  
Computer Aided ◽  
Energy Feature ◽  
Artificial Neural Network Ann ◽  
Ann Classifier ◽  
Occurrence Matrix ◽  
Aided Diagnosis

Computer-aided diagnosis system plays an important role in diagnosis and detection of breast cancer. In computer-aided diagnosis, feature extraction is one of the important steps. In this paper, we have proposed a method based on curvelet transform to classify mammogram images as normal -abnormal, benign and malignant. The feature vector is computed from the approximation coefficients. Directional energy is also calculated for all sub-bands. To select the efficient feature we used t-test and f-test methods. The selected feature is applied to Artificial Neural Network (ANN) classifier for classification. The effectiveness of the proposed method has been tested on MIAS database. The performance measures are computed with respect to normal vs. abnormal and benign vs. malignant for using approximation subband and energy feature of all curvelet coefficients. The highest classification accuracy of 95.34% is achieved for normal vs. abnormal and 80.86% is achieved for benign vs malignant class using energy feature of all curvelet coefficients.

scholarly journals 2D Convolutional Neural Markov Models for Spatiotemporal Sequence Forecasting

Sensors ◽  
2020 ◽  
Vol 20 (15) ◽  
pp. 4195
Author(s):  
Calvin Janitra Halim ◽  
Kazuhiko Kawamoto
Keyword(s):  
Neural Networks ◽  
Real World ◽  
Short Term Memory ◽  
Markov Models ◽  
Synthetic Data ◽  
Future Research ◽  
Target Sequence ◽  
Temporal Models ◽  
Spatial Aspect ◽  
Future Work

Recent approaches to time series forecasting, especially forecasting spatiotemporal sequences, have leveraged the approximation power of deep neural networks to model the complexity of such sequences, specifically approaches that are based on recurrent neural networks. Still, as spatiotemporal sequences that arise in the real world are noisy and chaotic, modeling approaches that utilize probabilistic temporal models, such as deep Markov models (DMMs), are favorable because of their ability to model uncertainty, increasing their robustness to noise. However, approaches based on DMMs do not maintain the spatial characteristics of spatiotemporal sequences, with most of the approaches converting the observed input into 1D data halfway through the model. To solve this, we propose a model that retains the spatial aspect of the target sequence with a DMM that consists of 2D convolutional neural networks. We then show the robustness of our method to data with large variance compared with naive forecast, vanilla DMM, and convolutional long short-term memory (LSTM) using synthetic data, even outperforming the DNN models over a longer forecast period. We also point out the limitations of our model when forecasting real-world precipitation data and the possible future work that can be done to address these limitations, along with additional future research potential.

Sign in / Sign up

Export Citation Format

Share Document