Neural Network Based Adult Image Classification

Kim, Won-Il; Lee, Hanku; Yoo, Seong Joon; Baik, Sung Wook

doi:10.1007/11550822_75

Cited by 14 publications

(14 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Artificial Neural Networkmentioning

confidence: 99%

“…Another interesting area of research is to employ the Artificial Neural Network (ANN) for image classification since its success greatly depends on the algorithm used in training ANNs [15]. ANNs are simple computational models that can handle different types of data.…”

Section: Introductionmentioning

confidence: 99%

“…In contrast, aside from color, Pounroja et al [2], Mhaski et al [13], and Fang et al [14] included other features such as texture, size, and shape in evaluating tomatoes in terms of quality and maturity. However, they differed in the machine learning techniques used to classify the tomato images into the different ripeness classes.Another interesting area of research is to employ the Artificial Neural Network (ANN) for image classification since its success greatly depends on the algorithm used in training ANNs [15]. ANNs are simple computational models that can handle different types of data.…”

mentioning

confidence: 99%

“…It is a nature-inspired computational paradigm that has interesting attributes like the ability to adapt, learn and generalize. Moreover, because of its massively parallel processing, ANNs are effective in solving many complex problems and have been extensively used in many areas such as in image segmentation and object classification problems [15,16,17,20,21,23,27,28,29].Also, artificial neural networks are computational models that have the ability to learn through adjustments of internal weight parameters as determined by the training algorithm in response to some training examples. The Multilayer Perceptron (MLP) is the most popular artificial neural network model used in many types of research.…”

mentioning

confidence: 99%

See 3 more Smart Citations

Automated Tomato Maturity Grading Using ABC-Trained Artificial Neural Networks

Opeña

Yusiong

2017

MJCS

View full text Add to dashboard Cite

Tomato maturity grading is quite essential in commercial farms that produce large quantities of tomatoes, and human graders usually perform tomato maturity grading. This task is carried out by matching the surface color of tomatoes to the United States Department of Agriculture's (USDA) tomato color chart which shows six maturity stages: Green, Breakers, Turning, Pink, Light Red, and Red. However, due to some uncontrollable factors, manual classification involving human graders is prone to misclassification. Thus, this paper introduces an automated tomato classification system that uses Artificial Neural Network (ANN) classifier trained using the Artificial Bee Colony (ABC) algorithm. To effectively classify tomatoes, the researchers combined five color features (Red, Green, Red-Green, Hue, a*) from three color models (RGB, HSI, CIE L*a*b*). These features are the inputs to the ANN classifier. Experiment results show that the ABC-trained ANN classifiers performed well in tomato classification and achieved high accuracy rate. Also, results show that combining the color features from different color models produce better accuracy rate than using color features from a single color model. With these results, an automated tomato classification system using an ABC-trained ANN classifier can be used to replace the manual classification procedure as it minimizes the chances of misclassification.Keywords: Tomato Grading, Artificial Bee Colony Algorithm, ABC, Artificial Neural Networks, Image Processing INTRODUCTIONTomato, Lycopersicon esculentum, is one of the most important and popular fruits in the world that can be consumed in several ways, for instance, as an ingredient in many dishes, sauces, and drinks. It has several varieties, which differ in shape, size, and color. Tomato maturity is commonly determined by its surface color because it is the most observable characteristic in estimating tomato maturity or ripeness, and most consumers buy tomatoes based only on the surface color. Hence, the color of the tomato is a major factor in determining the appropriate time to market the product since it influences consumers' preference [1,2,3].Fruits such as tomatoes are sorted and classified into several classes before being transported to commercial establishments for public consumption. This task is manually done using human labor. Farmers (human graders) in most commercial farms classify tomato maturity by comparing its surface color to the USDA color chart. The United States Department of Agriculture designed a color chart as a standard for tomato classification, and according to this chart, tomato ripeness is divided into six maturity stages: Green stage, Breakers stage, Turning stage, Pink stage, Light Red stage, and Red stage [1,4,5,7,8,17].However, the traditional way of tomato ripeness classification is very tiring and prone to error. Human graders can suffer from visual fatigue and stress that may affect the quality of grading. Also, human grading may vary from one person to another [3]. Because of these ...

show abstract

Section: Artificial Neural Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Automated Tomato Maturity Grading Using ABC-Trained Artificial Neural Networks

Opeña

Yusiong

2017

MJCS

View full text Add to dashboard Cite

show abstract

“…There are some investigators who carried out papers about adult image detection as: Xiaoyin Wang et al [6] who proposed an algorithm to detect adult images, Yue Wang et al [7] who proposed a way to help the algorithms to detect objectionable images using nipple detection, Huicheng Zheng et al [8] designed a filtering system to adult images, Wonil Kim et al [9] design a neural network based adult image classification, Jiann-Shu Lee et al [10]proposed an algorithm to naked image detection based on adaptive and extensible skin color model.…”

Section: Introductionmentioning

confidence: 99%