scholarly journals Vehicle Logo Recognition with Small Sample Problem in Complex Scene Based on Data Augmentation

2020 ◽  
Vol 2020 ◽  
pp. 1-10
Author(s):  
Xiao Ke ◽  
Pengqiang Du
Keyword(s):  
Gaussian Distribution ◽  
Data Augmentation ◽  
Small Sample Size ◽  
Small Sample ◽  
Segmentation Method ◽  
Complex Scene ◽  
Logo Recognition ◽  
Small Frame ◽  
Complex Scenes ◽  
Frame Method

Automatic identification for vehicles is an important topic in the field of Intelligent Transportation Systems (ITS), and the vehicle logo is one of the most important characteristics of a vehicle. Therefore, vehicle logo detection and recognition are important research topics. Because of the problems that the area of a vehicle logo is too small to be detected and the dataset is too small to train for complex scenes, considering the speed of recognition and the robustness to complex scenes, we use deep learning methods which are based on data optimization for vehicle logo in complex scenes. We propose three augmentation strategies for vehicle logo data: cross-sliding segmentation method, small frame method, and Gaussian Distribution Segmentation method. For the problem of small sample size, we use cross-sliding segmentation method, which can effectively increase the amount of data without changing the aspect ratio of the original vehicle logo image. To expand the area of the logos in the images, we develop the small frame method which improves the detection results of the small area vehicle logos. In order to enrich the position diversity of vehicle logo in the image, we propose Gaussian Distribution Segmentation method, and the result shows that this method is very effective. The F1 value of our method in the YOLO framework is 0.7765, and the precision is greatly improved to 0.9295. In the Faster R-CNN framework, the F1 value of our method is 0.7799, which is also better than before. The results of experiments show that the above optimization methods can better represent the features of the vehicle logos than the traditional method, and the experimental results have been improved.

scholarly journals Rectified Wing Loss for Efficient and Robust Facial Landmark Localisation with Convolutional Neural Networks

2019 ◽  
Vol 128 (8-9) ◽  
pp. 2126-2145 ◽  
Author(s):  
Zhen-Hua Feng ◽  
Josef Kittler ◽  
Muhammad Awais ◽  
Xiao-Jun Wu
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Loss Function ◽  
Data Augmentation ◽  
Small Sample Size ◽  
Small Sample ◽  
Imbalance Problem ◽  
Facial Landmark ◽  
The Impact ◽  
Coarse To Fine

AbstractEfficient and robust facial landmark localisation is crucial for the deployment of real-time face analysis systems. This paper presents a new loss function, namely Rectified Wing (RWing) loss, for regression-based facial landmark localisation with Convolutional Neural Networks (CNNs). We first systemically analyse different loss functions, including L2, L1 and smooth L1. The analysis suggests that the training of a network should pay more attention to small-medium errors. Motivated by this finding, we design a piece-wise loss that amplifies the impact of the samples with small-medium errors. Besides, we rectify the loss function for very small errors to mitigate the impact of inaccuracy of manual annotation. The use of our RWing loss boosts the performance significantly for regression-based CNNs in facial landmarking, especially for lightweight network architectures. To address the problem of under-representation of samples with large pose variations, we propose a simple but effective boosting strategy, referred to as pose-based data balancing. In particular, we deal with the data imbalance problem by duplicating the minority training samples and perturbing them by injecting random image rotation, bounding box translation and other data augmentation strategies. Last, the proposed approach is extended to create a coarse-to-fine framework for robust and efficient landmark localisation. Moreover, the proposed coarse-to-fine framework is able to deal with the small sample size problem effectively. The experimental results obtained on several well-known benchmarking datasets demonstrate the merits of our RWing loss and prove the superiority of the proposed method over the state-of-the-art approaches.

scholarly journals Approximated fast estimator for the shape parameter of generalized Gaussian distribution for a small sample size

2015 ◽  
Vol 63 (2) ◽  
pp. 405-411 ◽  
Author(s):  
R. Krupiński
Keyword(s):  
Sample Size ◽  
Gaussian Distribution ◽  
Shape Parameter ◽  
Small Sample Size ◽  
Estimation Method ◽  
Small Sample ◽  
Time Data ◽  
Generalized Gaussian Distribution ◽  
Root Finding ◽  
Real Time Data Processing

Abstract Most estimators of the shape parameter of generalized Gaussian distribution (GGD) assume asymptotic case when there is available infinite number of observations, but in the real case, there is only available a set of limited size. The most popular estimator for the shape parameter, i.e., the maximum likelihood (ML) method, has a larger variance with a decreasing sample size. A very high value of variance for a very small sample size makes this estimation method very inaccurate. A new fast approximated method based on the standardized moment to overcome this limitation is introduced in the article. The relative mean square error (RMSE) was plotted for the range 0.3-3 of the shape parameter for comparison with other methods. The method does not require any root finding, any long look-up table or multi step approach, therefore it is suitable for real-time data processing

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.

scholarly journals Res-WGAN: Image Classification for Plant Small-scale Datasets

2020 ◽  
Author(s):  
Mi Jiaqi ◽  
Hao xia ◽  
Yang Si ◽  
Yang Xiande ◽  
Gao Wanlin ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
User Study ◽  
Data Augmentation ◽  
Small Sample Size ◽  
Rare Plants ◽  
Small Sample ◽  
Primary Data ◽  
Small Scale ◽  
Rare Plant

Abstract Background: Artificial identification of rare plants is an important yet challenging problem in plant taxonomy. Although deep learning-based method can accurately predict rare plant category from training samples, accuracy requirements of only few experts are satisfied due to the small sample size. Thus, effective data augmentation is vital to improve the expressive power and robustness of deep learning models, especially for plant small-sample classification tasks. Different from traditional methods, a generative adversarial network (GAN) can mimic the distribution of primary data and produce similar but nonidentical samples. Data augmentation for automated classification of rare plant samples with GAN has not been studied for a long time. Result: In this study, we present a novel GAN model, referred to as residual Wasserstein GAN (Res-WGAN), for extending rare plant dataset efficiently. With a premise of minimized size of training parameters, residual block serves as elementary building block to increase the model’s expression ability. A loss function based on Wasserstein distance with gradient penalty is used to ensure diversity of output. Moreover, we draw on the super-resolution GAN (SRGAN) and consider perceptual loss into the function. Perceptual loss guarantees similarity between generated samples and original samples in high-dimensional features. Benefiting from these improvements, the expanded dataset improves the classification accuracy on the residual neural network (ResNet), a classic convolutional neural network, and restrains the overfitting phenomenon effectively by using transfer learning. More similar experiments are presented in this paper to prove the practicality of the model. Conclusions: Our proposed method produces better diversity and sharpness of generated samples and more accurate classification results than other data augmentation methods. In addition, a user study confirms that it is an ideal alternative strategy for small-scale rare plant identification. Developing a robust and effective small-scale plant classification method to replace expert testimony is highly relevant for agricultural automation development.

scholarly journals Improved Shrinkage Estimator of Large-Dimensional Covariance Matrix under the Complex Gaussian Distribution

2020 ◽  
Vol 2020 ◽  
pp. 1-8
Author(s):  
Bin Zhang
Keyword(s):  
Sample Size ◽  
Covariance Matrix ◽  
Gaussian Distribution ◽  
Array Signal Processing ◽  
Mean Squared Error ◽  
Small Sample Size ◽  
Linear Shrinkage ◽  
Small Sample ◽  
Shrinkage Estimator ◽  
Large Dimension

Estimating the covariance matrix of a random vector is essential and challenging in large dimension and small sample size scenarios. The purpose of this paper is to produce an outperformed large-dimensional covariance matrix estimator in the complex domain via the linear shrinkage regularization. Firstly, we develop a necessary moment property of the complex Wishart distribution. Secondly, by minimizing the mean squared error between the real covariance matrix and its shrinkage estimator, we obtain the optimal shrinkage intensity in a closed form for the spherical target matrix under the complex Gaussian distribution. Thirdly, we propose a newly available shrinkage estimator by unbiasedly estimating the unknown scalars involved in the optimal shrinkage intensity. Both the numerical simulations and an example application to array signal processing reveal that the proposed covariance matrix estimator performs well in large dimension and small sample size scenarios.

Probing the ultrastructure of the mitotic and meiotic spindle in eggs: An expeditious approach based on semi-thin (0.25 μm) serial sections

1983 ◽  
Vol 41 ◽  
pp. 552-555
Author(s):  
Conly L. Rieder ◽  
S. Bowser ◽  
R. Nowogrodzki ◽  
K. Ross ◽  
G. Sluder
Keyword(s):  
Small Sample Size ◽  
Small Sample ◽  
Meiotic Spindle ◽  
Serial Sections ◽  
Mitotic Apparatus ◽  
Thin Sections ◽  
Cell Cycles ◽  
Ultrastructural Studies

Eggs have long been a favorite material for studying the mechanism of karyokinesis in-vivo and in-vitro. They can be obtained in great numbers and, when fertilized, divide synchronously over many cell cycles. However, they are not considered to be a practical system for ultrastructural studies on the mitotic apparatus (MA) for several reasons, the most obvious of which is that sectioning them is a formidable task: over 1000 ultra-thin sections need to be cut from a single 80-100 μm diameter egg and of these sections only a small percentage will contain the area or structure of interest. Thus it is difficult and time consuming to obtain reliable ultrastructural data concerning the MA of eggs; and when it is obtained it is necessarily based on a small sample size.We have recently developed a procedure which will facilitate many studies concerned with the ultrastructure of the MA in eggs. It is based on the availability of biological HVEM's and on the observation that 0.25 μm thick serial sections can be screened at high resolution for content (after mounting on slot grids and staining with uranyl and lead) by phase contrast light microscopy (LM; Figs 1-2).

Validating the Factor Structure of the Stigma of Suicide Scale–Short Form Spanish Version Among Healthcare Students

Crisis ◽  
2020 ◽  
pp. 1-5
Author(s):  
Ruthmarie Hernández-Torres ◽  
Paola Carminelli-Corretjer ◽  
Nelmit Tollinchi-Natali ◽  
Ernesto Rosario-Hernández ◽  
Yovanska Duarté-Vélez ◽  
...  
Keyword(s):  
Factor Structure ◽  
Structural Equation ◽  
Short Form ◽  
Small Sample Size ◽  
Small Sample ◽  
Spanish Version ◽  
Equation Modeling ◽  
Cross Sectional ◽  
Healthcare Students ◽  
Spanish Speaking

Abstract. Background: Suicide is a leading cause of death among Spanish-speaking individuals. Suicide stigma can be a risk factor for suicide. A widely used measure is the Stigma of Suicide Scale-Short Form (SOSS-SF; Batterham, Calear, & Christensen, 2013 ). Although the SOSS-SF has established psychometric properties and factor structure in other languages and cultural contexts, no evidence is available from Spanish-speaking populations. Aim: This study aims to validate a Spanish translation of the SOSS-SF among a sample of Spanish-speaking healthcare students ( N = 277). Method: We implemented a cross-sectional design with quantitative techniques. Results: Following a structural equation modeling approach, a confirmatory factor analysis (CFA) supported the three-factor model proposed by Batterham and colleagues (2013) . Limitations: The study was limited by the small sample size and recruitment by availability. Conclusion: Findings suggest that the Spanish version of the SOSS-SF is a valid and reliable tool with which to examine suicide stigma among Spanish-speaking populations.

An Initial Investigation of Suicide Attempt Disclosures Among US Veterans

Crisis ◽  
2020 ◽  
pp. 1-7
Author(s):  
Brooke A. Ammerman ◽  
Sarah P. Carter ◽  
Heather M. Gebhardt ◽  
Jonathan Buchholz ◽  
Mark A. Reger
Keyword(s):  
Suicide Attempt ◽  
Mental Health Professionals ◽  
Help Seeking ◽  
Psychiatric Treatment ◽  
Small Sample Size ◽  
Romantic Partners ◽  
Small Sample ◽  
Professional Care ◽  
Patient Disclosure ◽  
Previous Suicide Attempt

Abstract. Background: Patient disclosure of prior suicidal behaviors is critical for effectively managing suicide risk; however, many attempts go undisclosed. Aims: The current study explored how responses following a suicide attempt disclosure may relate to help-seeking outcomes. Method: Participants included 37 veterans with a previous suicide attempt receiving inpatient psychiatric treatment. Veterans reported on their most and least helpful experiences disclosing their suicide attempt to others. Results: Veterans disclosed their suicide attempt to approximately eight individuals. Mental health professionals were the most cited recipient of their most helpful disclosure; romantic partners were the most common recipient of their least helpful disclosures. Positive reactions within the context of the least helpful disclosure experience were positively associated with a sense of connection with the disclosure recipient. Positive reactions within the most helpful disclosure experience were positively associated with the likelihood of future disclosure. No reactions were associated with having sought professional care or likelihood of seeking professional care. Limitations: The results are considered preliminary due to the small sample size. Conclusion: Findings suggest that while positive reactions may influence suicide attempt disclosure experiences broadly, additional research is needed to clarify factors that drive the decision to disclose a suicide attempt to a professional.

Investigating the Dynamics of Suicidal Ideation

Crisis ◽  
2018 ◽  
Vol 39 (1) ◽  
pp. 65-69 ◽  
Author(s):  
Nina Hallensleben ◽  
Lena Spangenberg ◽  
Thomas Forkmann ◽  
Dajana Rath ◽  
Ulrich Hegerl ◽  
...  
Keyword(s):  
Suicidal Ideation ◽  
Small Sample Size ◽  
Unipolar Depression ◽  
Small Sample ◽  
History Of ◽  
Severity Of Depression ◽  
Ecological Momentary ◽  
Depressive Episodes ◽  
Ema Method ◽  
Shed Light

Abstract. Background: Although the fluctuating nature of suicidal ideation (SI) has been described previously, longitudinal studies investigating the dynamics of SI are scarce. Aim: To demonstrate the fluctuation of SI across 6 days and up to 60 measurement points using smartphone-based ecological momentary assessments (EMA). Method: Twenty inpatients with unipolar depression and current and/or lifetime suicidal ideation rated their momentary SI 10 times per day over a 6-day period. Mean squared successive difference (MSSD) was calculated as a measure of variability. Correlations of MSSD with severity of depression, number of previous depressive episodes, and history of suicidal behavior were examined. Results: Individual trajectories of SI are shown to illustrate fluctuation. MSSD values ranged from 0.2 to 21.7. No significant correlations of MSSD with several clinical parameters were found, but there are hints of associations between fluctuation of SI and severity of depression and suicidality. Limitations: Main limitation of this study is the small sample size leading to low power and probably missing potential effects. Further research with larger samples is necessary to shed light on the dynamics of SI. Conclusion: The results illustrate the dynamic nature and the diversity of trajectories of SI across 6 days in psychiatric inpatients with unipolar depression. Prediction of the fluctuation of SI might be of high clinical relevance. Further research using EMA and sophisticated analyses with larger samples is necessary to shed light on the dynamics of SI.

Sign in / Sign up

Export Citation Format

Share Document