Deep Learning Based Tobacco Products Classification (Preprint)

2019 ◽  
Author(s):  
Murat Taskiran ◽  
Sibel Cimen Yetis
Keyword(s):  
Young People ◽  
Feature Vector ◽  
Weighted Average ◽  
Classification Performance ◽  
Accuracy Rate ◽  
Tobacco Products ◽  
Internet Environment ◽  
Performance Rate ◽  
Fully Connected

BACKGROUND Various images and videos are uploaded every day or even every second on Instagram. These publicly available images are easily accessible as a result of uncontrolled Internet use by young people and children. Shared images include tobacco products and can be encouraging for young people and children when they are accessible. OBJECTIVE In this study, it is aimed to detect tobacco and tobacco products with various Convolutional Neural Networks (CNNs). METHODS A total of 1607 public images were collected from Instagram, and feature vectors were extracted with various CNNs, which proved to be successful in the competitions and CNN was determined to be proper for this problem. RESULTS MobileNet gave the highest results 99.1% as weighted average. The feature vector of the input images are extracted with CNNs and classified with the latest fully connected layer. CONCLUSIONS The classification of the tobacco products of 4 different classes was studied by using the networks and the classification performance rate was obtained as 100% for 322 test images via MobileNet. In this way, the content that is encouraging for children can be censored or filtered with a high accuracy rate and a secure Internet environment can be provided.

Deep Learning Based Tobacco Products Classification (Preprint)

2019 ◽  
Author(s):  
Murat Taşkıran ◽  
Sibel Çimen Yetiş
Keyword(s):  
Young People ◽  
Feature Vector ◽  
Weighted Average ◽  
Classification Performance ◽  
Accuracy Rate ◽  
Tobacco Products ◽  
Internet Environment ◽  
Performance Rate ◽  
Fully Connected

BACKGROUND Various images and videos are uploaded every day or even every second on Instagram. These publicly available images are easily accessible as a result of uncontrolled Internet use by young people and children. Shared images include tobacco products and can be encouraging for young people and children when they are accessible. OBJECTIVE In this study, it is aimed to detect tobacco and tobacco products with various Convolutional Neural Networks (CNNs) and to limit the access of young users to these detected tobacco products over the Internet. METHODS A total of 1607 public images were collected from Instagram, and feature vectors were extracted with various CNNs, which proved to be successful in the competitions and CNN was determined to be proper for detect tobacco products. RESULTS MobileNet gave the highest results 99.1% as weighted average. The feature vector of the input images are extracted with CNNs and classified with the latest fully connected layer. CONCLUSIONS The classification of the tobacco products of 4 different classes was studied by using the networks and the classification performance rate was obtained as 100% for 322 test images via MobileNet. In this way, the content that is encouraging for children can be censored or filtered with a high accuracy rate and a secure Internet environment can be provided.

Wavelet Transform Based Feature Extraction for EEG Signal Classification

2021 ◽  
Vol 20 ◽  
pp. 199-206
Author(s):  
Seda Postalcioglu
Keyword(s):  
Wavelet Transform ◽  
High Performance ◽  
Feature Vector ◽  
Fast Response ◽  
Eeg Signal ◽  
Real Time Control ◽  
Time Control ◽  
Band Power ◽  
Performance Rate

This study focused on the classification of EEG signal. The study aims to make a classification with fast response and high-performance rate. Thus, it could be possible for real-time control applications as Brain-Computer Interface (BCI) systems. The feature vector is created by Wavelet transform and statistical calculations. It is trained and tested with a neural network. The db4 wavelet is used in the study. Pwelch, skewness, kurtosis, band power, median, standard deviation, min, max, energy, entropy are used to make the wavelet coefficients meaningful. The performance is achieved as 99.414% with the running time of 0.0209 seconds

Feature Modeling and Classification of Halftone Image Based on Statistical Method

2012 ◽  
Vol 229-231 ◽  
pp. 1693-1696
Author(s):  
Zhi Qiang Wen ◽  
Wen Qiu Zhu ◽  
Yong Xiang Hu ◽  
Zhao Yi Peng
Keyword(s):  
Autocorrelation Function ◽  
Feature Vector ◽  
Classification Performance ◽  
Feature Modeling ◽  
Gray Level ◽  
Halftone Image ◽  
Occurrence Matrix ◽  
Halftone Images ◽  
Better Than

For problem of feature modeling on halftone image, three statistics methods, named gray-level co-occurrence matrix, autocorrelation function and spectrum statistics, are used to extract feature vector of various halftone images. Then, their classification performance is assessed by radial basis function neural network. A mass of experiments show the autocorrelation function is better than other two methods for classification on halftone image.

scholarly journals A study on the effects of diffusion of information on epidemic spread

2019 ◽  
Vol 10 (03) ◽  
pp. 1950015 ◽  
Author(s):  
Semra Gündüç
Keyword(s):  
Direct Contact ◽  
Information Diffusion ◽  
Weighted Average ◽  
Scale Free ◽  
Diffusion Of Information ◽  
Scale Free Networks ◽  
Underlying Network ◽  
Fully Connected ◽  
Taking Action

In this work, the spread of a contagious disease on a society where the individuals may take precautions is modeled. The primary assumption is that the infected individuals transmit the infection to the susceptible members of the community through direct contact interactions. In the meantime, the susceptibles gather information from the adjacent sites which may lead to taking precautions. The SIR model is used for the diffusion of infection while the Bass equation models the information diffusion. The sociological classification of the individuals indicates that a small percentage of the population takes action immediately after being informed, while the majority expects to see some real advantage of taking action. The individuals are assumed to take two different precautions. The precursory measures are getting vaccinated or trying to avoid direct contact with the neighbors. A weighted average of states of the neighbors leads to the choice of action. The fully connected and scale-free Networks are employed as the underlying network of interactions. The comparison between the simple contagion diffusion and the diffusion of infection in a responsive society showed that a very limited precaution makes a considerable difference in the speed and the size of the spread of illness. Particularly, highly connected hub nodes play an essential role in the reduction of the spread of disease.

scholarly journals Selective Feature Generation Method for Classification of Low-Dimensional Data

2018 ◽  
Vol 13 (1) ◽  
pp. 24
Author(s):  
Sang-Il Choi ◽  
Sang Tae Choi ◽  
Haanju Yoo
Keyword(s):  
Feature Vector ◽  
Image Data ◽  
Classification Performance ◽  
Data Sets ◽  
Facial Image ◽  
Discrimination Power ◽  
Input Feature ◽  
Low Dimensional ◽  
Candidate Feature

We propose a method that generates input features to effectively classify low-dimensional data. To do this, we first generate high-order terms for the input features of the original low-dimensional data to form a candidate set of new input features. Then, the discrimination power of the candidate input features is quantitatively evaluated by calculating the ‘discrimination distance’ for each candidate feature. As a result, only candidates with a large amount of discriminative information are selected to create a new input feature vector, and the discriminant features that are to be used as input to the classifier are extracted from the new input feature vectors by using a subspace discriminant analysis. Experiments on low-dimensional data sets in the UCI machine learning repository and several kinds of low-resolution facial image data show that the proposed method improves the classification performance of low-dimensional data by generating features.

scholarly journals An Efficient Machine Learning Method for Facial Expression Recognition

2019 ◽  
Vol 8 (9S) ◽  
pp. 539-546
Keyword(s):  
Feature Vector ◽  
Rejection Rate ◽  
Extraction Process ◽  
Expression Recognition ◽  
Accuracy Rate ◽  
Efficient Machine ◽  
Feature Optimization ◽  
Error Probabilities ◽  
Intelligence Approach

Emotions play important role in human sentiments so broad studies are carried out to explore the relation between human sentiments and machine interactions. This paper deals with an automatic system which spontaneously identifies the facial emotion. Gradient filtering and component analysis is used to extract feature vector and feature optimization is taken place using swarm intelligence approach. Thus emotion recognition with optimized feature extraction process is carried out with high accuracy rate and less error probabilities. Finally the testing process is obtained for the classification of emotions and then performance is measured in terms of false acceptance rate, false rejection rate, and accuracy.

A Study on the Auxiliary Diagnosis of Thyroid Disease Images Based on Multiple Dimensional Deep Learning Algorithms

2020 ◽  
Vol 16 (3) ◽  
pp. 199-205
Author(s):  
Yuejun Liu ◽  
Yifei Xu ◽  
Xiangzheng Meng ◽  
Xuguang Wang ◽  
Tianxu Bai
Keyword(s):  
Deep Learning ◽  
Learning Algorithms ◽  
Region Of Interest ◽  
Classification Performance ◽  
Thyroid Diseases ◽  
Great Success ◽  
Learning Models ◽  
Good Classification Performance ◽  
Spect Images

Background: Medical imaging plays an important role in the diagnosis of thyroid diseases. In the field of machine learning, multiple dimensional deep learning algorithms are widely used in image classification and recognition, and have achieved great success. Objective: The method based on multiple dimensional deep learning is employed for the auxiliary diagnosis of thyroid diseases based on SPECT images. The performances of different deep learning models are evaluated and compared. Methods: Thyroid SPECT images are collected with three types, they are hyperthyroidism, normal and hypothyroidism. In the pre-processing, the region of interest of thyroid is segmented and the amount of data sample is expanded. Four CNN models, including CNN, Inception, VGG16 and RNN, are used to evaluate deep learning methods. Results: Deep learning based methods have good classification performance, the accuracy is 92.9%-96.2%, AUC is 97.8%-99.6%. VGG16 model has the best performance, the accuracy is 96.2% and AUC is 99.6%. Especially, the VGG16 model with a changing learning rate works best. Conclusion: The standard CNN, Inception, VGG16, and RNN four deep learning models are efficient for the classification of thyroid diseases with SPECT images. The accuracy of the assisted diagnostic method based on deep learning is higher than that of other methods reported in the literature.

scholarly journals Freely scalable and reconfigurable optical hardware for deep learning

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Liane Bernstein ◽  
Alexander Sludds ◽  
Ryan Hamerly ◽  
Vivienne Sze ◽  
Joel Emer ◽  
...  
Keyword(s):  
Neural Network ◽  
Energy Consumption ◽  
Data Transfer ◽  
Power Delivery ◽  
Optical Data ◽  
Optical Multicast ◽  
Optical Neural Network ◽  
Fully Connected ◽  
Management Power

AbstractAs deep neural network (DNN) models grow ever-larger, they can achieve higher accuracy and solve more complex problems. This trend has been enabled by an increase in available compute power; however, efforts to continue to scale electronic processors are impeded by the costs of communication, thermal management, power delivery and clocking. To improve scalability, we propose a digital optical neural network (DONN) with intralayer optical interconnects and reconfigurable input values. The path-length-independence of optical energy consumption enables information locality between a transmitter and a large number of arbitrarily arranged receivers, which allows greater flexibility in architecture design to circumvent scaling limitations. In a proof-of-concept experiment, we demonstrate optical multicast in the classification of 500 MNIST images with a 3-layer, fully-connected network. We also analyze the energy consumption of the DONN and find that digital optical data transfer is beneficial over electronics when the spacing of computational units is on the order of $$>10\,\upmu $$ > 10 μ m.

Explainable Fully Connected Visual Words for the Classification of Skin Cancer Confocal Images

2020 ◽  
Author(s):  
Athanasios Kallipolitis ◽  
Alexandros Stratigos ◽  
Alexios Zarras ◽  
Ilias Maglogiannis
Keyword(s):  
Skin Cancer ◽  
Visual Words ◽  
Confocal Images ◽  
Fully Connected
Sign in / Sign up

Export Citation Format

Share Document