scholarly journals Micro-expression recognition with small sample size by transferring long-term convolutional neural network

2018 ◽  
Vol 312 ◽  
pp. 251-262 ◽  
Author(s):  
Su-Jing Wang ◽  
Bing-Jun Li ◽  
Yong-Jin Liu ◽  
Wen-Jing Yan ◽  
Xinyu Ou ◽  
...  
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Sample Size ◽  
Small Sample Size ◽  
Small Sample ◽  
Expression Recognition ◽  
Micro Expression

scholarly journals Identification of cucumber leaf diseases using deep learning and small sample size for agricultural Internet of Things

2021 ◽  
Vol 17 (4) ◽  
pp. 155014772110074
Author(s):  
Jingyao Zhang ◽  
Yuan Rao ◽  
Chao Man ◽  
Zhaohui Jiang ◽  
Shaowen Li
Keyword(s):  
Neural Network ◽  
Internet Of Things ◽  
Convolutional Neural Network ◽  
Sample Size ◽  
Data Augmentation ◽  
Small Sample Size ◽  
Small Sample ◽  
Identification Accuracy ◽  
Generative Adversarial Networks ◽  
Training Samples

Due to the complex environments in real fields, it is challenging to conduct identification modeling and diagnosis of plant leaf diseases by directly utilizing in-situ images from the system of agricultural Internet of things. To overcome this shortcoming, one approach, based on small sample size and deep convolutional neural network, was proposed for conducting the recognition of cucumber leaf diseases under field conditions. One two-stage segmentation method was presented to acquire the lesion images by extracting the disease spots from cucumber leaves. Subsequently, after implementing rotation and translation, the lesion images were fed into the activation reconstruction generative adversarial networks for data augmentation to generate new training samples. Finally, to improve the identification accuracy of cucumber leaf diseases, we proposed dilated and inception convolutional neural network that was trained using the generated training samples. Experimental results showed that the proposed approach achieved the average identification accuracy of 96.11% and 90.67% when implemented on the data sets of lesion and raw field diseased leaf images with three different diseases of anthracnose, downy mildew, and powdery mildew, significantly outperforming those existing counterparts, indicating that it offered good potential of serving field application of agricultural Internet of things.

Micro-expression recognition based on 3D flow convolutional neural network

2018 ◽  
Vol 22 (4) ◽  
pp. 1331-1339 ◽  
Author(s):  
Jing Li ◽  
Yandan Wang ◽  
John See ◽  
Wenbin Liu
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Expression Recognition ◽  
3D Flow ◽  
Micro Expression

scholarly journals Surgically accelerated orthodontic techniques and periodontal response: a systematic review

2020 ◽  
Vol 42 (6) ◽  
pp. 635-642 ◽  
Author(s):  
Umar Rekhi ◽  
Raisa Queiroz Catunda ◽  
Monica Prasad Gibson
Keyword(s):  
Systematic Review ◽  
High Risk ◽  
Sample Size ◽  
Orthodontic Treatment ◽  
Small Sample Size ◽  
Small Sample ◽  
Risk Of Bias ◽  
Long Term Follow Up

Summary Background Reduction in orthodontic treatment time is gaining popularity due to patient demands. Several new techniques of acceleratory orthodontic treatment have been introduced to effectively treat the malocclusion in a shorter time period with minimal adverse effects. Objective The objective of this systematic review is to critically evaluate the potential effect of accelerated surgically assisted orthodontic techniques on periodontal tissues. Materials and methods Electronic databases used to perform the search were Medline (Ovid), EMBASE, PubMed, Scopus, Cochrane, Google Scholar, and hand searching of the literature was also performed. Selection criteria Only randomized control trials (RCTs) that assessed the relationship between accelerated surgically assisted orthodontic techniques and its effects on periodontium were included. Data collection and analysis The Joanna Briggs Institute (JBI) critical appraisal checklist tool (2016) was used to assess the finally selected studies. Among these studies, five evaluated corticotomy-facilitated orthodontics, two tested accelerated tooth movement with piezocision, one compared corticotomy-facilitated orthodontics with piezocision, and one studied the effects of periodontally accelerated osteogenic orthodontics. The duration of these studies was relatively short and had moderate to high risk of bias. Results Literature search identified 225 records from 5 databases and 50 articles from the partial grey literature (Google scholar) search. Finally, nine eligible RCTs were included in the review. Limitations Most of the included studies were of a high risk of bias due to high experimental heterogeneity and small sample size. Long-term follow-up of the periodontal response to these interventions was also lacking. Conclusions There is an absence of evidence considering the lack of long-term follow-up and small sample size therefore, the results of this review should be carefully interpreted. Implications Due to the need for more studies with less risk of bias, these techniques should be implemented in dental practice with caution. With stronger evidence, the study may be confirmed to provide quicker desired results for orthodontic patients. Registration This study protocol was not registered. Funding No funding was obtained for this systematic review.

Non-equidistant GM(1,1) model based on GCHM_WBO and its application in corrosion rate prediction

2016 ◽  
Vol 6 (3) ◽  
pp. 365-374 ◽  
Author(s):  
Yuanjie Zhi ◽  
Dongmei Fu ◽  
Hanling Wang
Keyword(s):  
Corrosion Rate ◽  
Sample Size ◽  
Prediction Accuracy ◽  
Small Sample Size ◽  
Real Data ◽  
Small Sample ◽  
Remaining Useful Life ◽  
Content Type ◽  
Corrosion Data

Purpose The purpose of this paper is to present a new model which combines the non-equidistant GM(1,1) model with GCHM_WBO (generalized contra-harmonic mean (GCHM); weakening buffer operator (WBO)). The authors use the model to solve the deadlock that for a large number of non-equidistant corrosion rate, it is difficult to establish a reasonable prediction model and improve the prediction accuracy. Design/methodology/approach This research consists of three parts: non-equidistant GM(1,1) model, GCHM_WBO operator, and the optimization of morphing parameter (contained in GCHM, control the intensity of the weakening operator). The methodology is explained as follows. First, the authors built a non-equidistant GM(1,1) model with GCHM_WBO weakened data, of which morphing parameter was randomly selected. Next, the authors calculated the error between prediction data of model and the real data, and adjusted the morphing parameter according to the error and property of GCHM. Then, the authors generated a new non-equidistant GM(1,1) based on new morphing parameter, and repeated the previous step until the termination condition was satisfied. Finally, the model with appropriate morphing parameter was used to implement the prediction of new data. Findings This paper finds a property of GCHM, which is a monotonic increasing function of morphing parameter in some specific conditions. Based on the property and the fixed point axiom of WBO, an algorithm was designed to search an appropriate morphing parameter. The appropriate morphing parameter was implemented for the purpose of improving the accuracy of the model. The model was applied to predict the corrosion rate of six steels at Guangzhou experimental station. The results showed that the proposed method can get more accuracy in prediction capability compared to the models with the original data and AWBO weakened data. The method is applicable to long-term forecasts in case of data scarcity. Practical implications Corrosion will cause huge economic loss to a country; therefore, it is important to judge the remaining useful life of a material or equipment; the foundation for judgement of which is the prediction of material corrosion rate. However, the prediction of corrosion rate is very difficult because of corrosion data’s features, such as small sample size, non-equidistant, etc. The proposed method can be used to implement long-term forecast of corrosion data with only one sample and non-equidistant samples. Originality/value This paper presented a model which combines the non-equidistant GM(1,1) model with GCHM_WBO to handle the problem of long-term forecasting of corrosion data. In the modelling process, the proposed morphing parameter searched through algorithm can improve the prediction accuracy of the model. Therefore, the model can provide effective and reliable result when data are of a small sample size and non-equidistant.

scholarly journals The CURAT Sonification Game: Gamification for Remote Sonification Evaluation

2021 ◽  
Author(s):  
Tim Ziemer ◽  
Holger Schultheis
Keyword(s):  
Sample Size ◽  
Experimental Evaluation ◽  
Small Sample Size ◽  
Large Population ◽  
Small Sample ◽  
Short Term ◽  
Long Period ◽  
Future Developments

As sonification is supposed to communicate information to users, experimental evaluation of the subjective appropriateness and effectiveness of the sonification design is often desired and sometimes indispensable. Experiments in the laboratory are typically restricted to short-term usage by a small sample size under unnatural conditions. We introduce the multi-platform CURAT Sonification Game that allows us to evaluate our sonification design by a large population during long-term usage. Gamification is used to motivate users to interact with the sonification regularly and conscientiously over a long period of time. In this paper we present the sonification game and some initial analyses of the gathered data. Furthermore, we hope to reach more volunteers to play the CURAT Sonification Game and help us evaluate and optimize our psychoacoustic sonification design and give us valuable feedback on the game and recommendations for future developments.

scholarly journals Spatiotemporal Convolutional Neural Network with Convolutional Block Attention Module for Micro-Expression Recognition

Information ◽  
2020 ◽  
Vol 11 (8) ◽  
pp. 380
Author(s):  
Boyu Chen ◽  
Zhihao Zhang ◽  
Nian Liu ◽  
Yang Tan ◽  
Xinyu Liu ◽  
...  
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Three Dimensional ◽  
Expression Recognition ◽  
Learning Framework ◽  
The Face ◽  
3D Cnn ◽  
Visual Attention Mechanism ◽  
Casme Ii ◽  
Micro Expression

A micro-expression is defined as an uncontrollable muscular movement shown on the face of humans when one is trying to conceal or repress his true emotions. Many researchers have applied the deep learning framework to micro-expression recognition in recent years. However, few have introduced the human visual attention mechanism to micro-expression recognition. In this study, we propose a three-dimensional (3D) spatiotemporal convolutional neural network with the convolutional block attention module (CBAM) for micro-expression recognition. First image sequences were input to a medium-sized convolutional neural network (CNN) to extract visual features. Afterwards, it learned to allocate the feature weights in an adaptive manner with the help of a convolutional block attention module. The method was testified in spontaneous micro-expression databases (Chinese Academy of Sciences Micro-expression II (CASME II), Spontaneous Micro-expression Database (SMIC)). The experimental results show that the 3D CNN with convolutional block attention module outperformed other algorithms in micro-expression recognition.

Sign in / Sign up

Export Citation Format

Share Document