A SEGMENTATION-FREE RECOGNITION OF HANDWRITTEN TOUCHING NUMERAL PAIRS USING MODULAR NEURAL NETWORK

2001 ◽  
Vol 15 (06) ◽  
pp. 949-966 ◽  
Author(s):  
SOON-MAN CHOI ◽  
Il-SEOK OH
Keyword(s):  
Neural Network ◽  
Conventional Approach ◽  
Experimental Results ◽  
Large Set ◽  
Training Set ◽  
Modular Neural Network ◽  
Large Size

The conventional approach to the recognition of handwritten touching numeral pairs uses a process with two steps; splitting the touching numerals and recognizing individual numerals. It shows a limitation mainly due to a large variation in touching styles between two numerals. In this paper, we adopt the segmentation-free approach, which regards a touching numeral pair as an atomic pattern. Two important issues are raised, i.e. solving the large-set classification and constructing a large-size training set. For the 100-class classification, we use a modular neural network which consists of 100 separate subnetworks. We construct the training set with a balance among 100 classes and using a sufficient amount by extracting actual samples from a numeral database and synthesizing samples with a scheme of forcing two numerals to touch. The experimental results show a promising performance.

scholarly journals The Recognition of Rice Area Images by UAV Based on Deep Learning

2018 ◽  
Vol 232 ◽  
pp. 02057 ◽  
Author(s):  
Hongyuan Wei ◽  
Jian Mao
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Feature Recognition ◽  
Rice Field ◽  
Training Model ◽  
Practical Application ◽  
Training Set ◽  
Large Size ◽  
Rice Area ◽  
Python Programming

Aiming at the target detection of remote sensing rice field of uav, the image of large-size uav is firstly segmented, and the type of each image is manually identified, and the image training set and verification set are made. Then, the training model of convolutional neural network is realized by python programming. The advantage and disadvantage of the two-layer convolutional neural network and ResNet50 are compared, and it is found that the training set is less and the picture feature complexity is not high in practical application. In the end, the feature recognition of rice field is realized, which has certain application value.

Barley Diseases Discrimination Based on Neural Network

2014 ◽  
Vol 687-691 ◽  
pp. 3914-3916
Author(s):  
Lin Ming Wang
Keyword(s):  
Neural Network ◽  
Image Recognition ◽  
New Technology ◽  
Texture Features ◽  
Experimental Results ◽  
Neural Net ◽  
Training Set ◽  
The Neural Network ◽  
Test Sets ◽  
Spot Color

First disease spot color and texture features were extracted from barley field images in Gansu, and the feature vectors were used as input vector to establish barley diseases classifier model. Then the neural network was applied to rain classified model with collected images as training set. Finally, two groups of random selected images as test sets were used to perform classified verification experiments. The experimental results show that the overall accuracy of barley dis-eases recognition model is above 86.7%. Therefore, Barley disease image recognition based on neural net-work provides a new technology for the classified treatment of barley diseases.

scholarly journals Order-Free Learning Alleviating Exposure Bias in Multi-Label Classification

2020 ◽  
Vol 34 (04) ◽  
pp. 6038-6045
Author(s):  
Che-Ping Tsai ◽  
Hung-Yi Lee
Keyword(s):  
Neural Network ◽  
Recurrent Neural Network ◽  
Experimental Results ◽  
Prediction Problem ◽  
Generalization Capability ◽  
Training Set ◽  
Large Margin ◽  
Sequence Prediction ◽  
Benchmark Datasets ◽  
New Framework

Multi-label classification (MLC) assigns multiple labels to each sample. Prior studies show that MLC can be transformed to a sequence prediction problem with a recurrent neural network (RNN) decoder to model the label dependency. However, training a RNN decoder requires a predefined order of labels, which is not directly available in the MLC specification. Besides, RNN thus trained tends to overfit the label combinations in the training set and have difficulty generating unseen label sequences. In this paper, we propose a new framework for MLC which does not rely on a predefined label order and thus alleviates exposure bias. The experimental results on three multi-label classification benchmark datasets show that our method outperforms competitive baselines by a large margin. We also find the proposed approach has a higher probability of generating label combinations not seen during training than the baseline models. The result shows that the proposed approach has better generalization capability.

Deep Learning Application for Plant Classification on Unbalanced Training Set

2019 ◽  
Author(s):  
Rafael S. Pereira ◽  
Fabio Porto
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
High Probability ◽  
Experimental Results ◽  
Learning Models ◽  
Classification Problems ◽  
Training Set ◽  
Species Classification ◽  
Accuracy Level ◽  
Plant Image

Deep learning models expect a reasonable amount of training in- stances to improve prediction quality. Moreover, in classification problems, the occurrence of an unbalanced distribution may lead to a biased model. In this paper, we investigate the problem of species classification from plant images, where some species have very few image samples. We explore reduced versions of imagenet Neural Network winners architecture to filter the space of candi- date matches, under a target accuracy level. We show through experimental results using real unbalanced plant image datasets that our approach can lead to classifications within the 5 best positions with high probability.

scholarly journals An Innovative Way to Measure the Quality of a Neural Network Without the Use of a Test Set

2001 ◽  
Vol 5 (1) ◽  
pp. 31-36 ◽  
Author(s):  
Giovanni Pilato ◽  
◽  
Filippo Sorbello ◽  
Giorgio Vassallo
Keyword(s):  
Neural Network ◽  
Computational Cost ◽  
Experimental Results ◽  
Output Function ◽  
Quality Factors ◽  
Generalization Capability ◽  
Training Set ◽  
Test Set ◽  
Network Output

In this paper, three quality factors are introduced in order to measure the quality of a neural network. Each factor deals with a particular feature of quality: the ability of the network in learning training set samples; generalization capability related to the gradient, in the nearby of the training patterns, of the network output function; the computational cost of the architecture during the production phase, related to the number of connections between neural units. The validity of the proposed solution has been tested using three well-known benchmarks. Experimental results show that quality factors introduced in this paper can be a valid alternative to the test set method.

scholarly journals Deep Learning Application for Plant Classification on Unbalanced Training Set

2019 ◽  
Author(s):  
Rafael S. Pereira ◽  
Fabio Porto
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
High Probability ◽  
Experimental Results ◽  
Learning Models ◽  
Classification Problems ◽  
Training Set ◽  
Species Classification ◽  
Accuracy Level ◽  
Plant Image

Deep learning models expect a reasonable amount of training instances to improve prediction quality. Moreover, in classification problems, the occurrence of an unbalanced distribution may lead to a biased model. In this paper, we investigate the problem of species classification from plant images, where some species have very few image samples. We explore reduced versions of imagenet Neural Network winners architecture to filter the space of candidate matches, under a target accuracy level. We show through experimental results using real unbalanced plant image datasets that our approach can lead to classifications within the 5 best positions with high probability.  

Adsorption Isotherm Predictions for Multiple Molecules in MOFs Using the Same Deep Learning Model

2019 ◽  
Author(s):  
Ryther Anderson ◽  
Achay Biong ◽  
Diego Gómez-Gualdrón
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Molecular Simulation ◽  
Large Scale ◽  
Learning Model ◽  
Operating Conditions ◽  
Small Subset ◽  
Screening Methods ◽  
Large Set ◽  
Metal Organic

<div>Tailoring the structure and chemistry of metal-organic frameworks (MOFs) enables the manipulation of their adsorption properties to suit specific energy and environmental applications. As there are millions of possible MOFs (with tens of thousands already synthesized), molecular simulation, such as grand canonical Monte Carlo (GCMC), has frequently been used to rapidly evaluate the adsorption performance of a large set of MOFs. This allows subsequent experiments to focus only on a small subset of the most promising MOFs. In many instances, however, even molecular simulation becomes prohibitively time consuming, underscoring the need for alternative screening methods, such as machine learning, to precede molecular simulation efforts. In this study, as a proof of concept, we trained a neural network as the first example of a machine learning model capable of predicting full adsorption isotherms of different molecules not included in the training of the model. To achieve this, we trained our neural network only on alchemical species, represented only by their geometry and force field parameters, and used this neural network to predict the loadings of real adsorbates. We focused on predicting room temperature adsorption of small (one- and two-atom) molecules relevant to chemical separations. Namely, argon, krypton, xenon, methane, ethane, and nitrogen. However, we also observed surprisingly promising predictions for more complex molecules, whose properties are outside the range spanned by the alchemical adsorbates. Prediction accuracies suitable for large-scale screening were achieved using simple MOF (e.g. geometric properties and chemical moieties), and adsorbate (e.g. forcefield parameters and geometry) descriptors. Our results illustrate a new philosophy of training that opens the path towards development of machine learning models that can predict the adsorption loading of any new adsorbate at any new operating conditions in any new MOF.</div>

scholarly journals Geometric property-based convolutional neural network for indoor object detection

2021 ◽  
Vol 18 (1) ◽  
pp. 172988142199332
Author(s):  
Xintao Ding ◽  
Boquan Li ◽  
Jinbao Wang
Keyword(s):  
Neural Network ◽  
Object Detection ◽  
Convolutional Neural Network ◽  
Geometric Property ◽  
Ground Truth ◽  
Geometric Constraints ◽  
Depth Information ◽  
Training Set ◽  
Object Knowledge ◽  
The Mean

Indoor object detection is a very demanding and important task for robot applications. Object knowledge, such as two-dimensional (2D) shape and depth information, may be helpful for detection. In this article, we focus on region-based convolutional neural network (CNN) detector and propose a geometric property-based Faster R-CNN method (GP-Faster) for indoor object detection. GP-Faster incorporates geometric property in Faster R-CNN to improve the detection performance. In detail, we first use mesh grids that are the intersections of direct and inverse proportion functions to generate appropriate anchors for indoor objects. After the anchors are regressed to the regions of interest produced by a region proposal network (RPN-RoIs), we then use 2D geometric constraints to refine the RPN-RoIs, in which the 2D constraint of every classification is a convex hull region enclosing the width and height coordinates of the ground-truth boxes on the training set. Comparison experiments are implemented on two indoor datasets SUN2012 and NYUv2. Since the depth information is available in NYUv2, we involve depth constraints in GP-Faster and propose 3D geometric property-based Faster R-CNN (DGP-Faster) on NYUv2. The experimental results show that both GP-Faster and DGP-Faster increase the performance of the mean average precision.

scholarly journals Artificial Neural Network Simulation of Energetic Performance for Sorption Thermal Energy Storage Reactors

Energies ◽  
2021 ◽  
Vol 14 (11) ◽  
pp. 3294
Author(s):  
Carla Delmarre ◽  
Marie-Anne Resmond ◽  
Frédéric Kuznik ◽  
Christian Obrecht ◽  
Bao Chen ◽  
...  
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Energy Storage ◽  
Thermal Energy ◽  
Temperature Difference ◽  
Thermal Energy Storage ◽  
Heat Storage ◽  
Experimental Results ◽  
Physical Models ◽  
Artificial Neural

Sorption thermal heat storage is a promising solution to improve the development of renewable energies and to promote a rational use of energy both for industry and households. These systems store thermal energy through physico-chemical sorption/desorption reactions that are also termed hydration/dehydration. Their introduction to the market requires to assess their energy performances, usually analysed by numerical simulation of the overall system. To address this, physical models are commonly developed and used. However, simulation based on such models are time-consuming which does not allow their use for yearly simulations. Artificial neural network (ANN)-based models, which are known for their computational efficiency, may overcome this issue. Therefore, the main objective of this study is to investigate the use of an ANN model to simulate a sorption heat storage system, instead of using a physical model. The neural network is trained using experimental results in order to evaluate this approach on actual systems. By using a recurrent neural network (RNN) and the Deep Learning Toolbox in MATLAB, a good accuracy is reached, and the predicted results are close to the experimental results. The root mean squared error for the prediction of the temperature difference during the thermal energy storage process is less than 3K for both hydration and dehydration, the maximal temperature difference being, respectively, about 90K and 40K.

Sign in / Sign up

Export Citation Format

Share Document