Data Augmentation Using Contour Image for Convolutional Neural Network

With the development of artificial intelligence-related technologies such as deep learning,various organizations,including the government,are making various efforts to generate and manage big data for use in artificial intelligence.However,it is difficult to acquire big data due to various social problems and restrictions such as personal information leakage.There are many problems in introducing technology in fields that do not have enough training data necessary to apply deep learning technology.Therefore,this study proposes a mixed contour data augmentation technique,which is a data augmentation technique using contour images,to solve a problem caused by a lack of data.ResNet,a famous convolutional neural network(CNN)architecture,and CIFAR-10,a benchmark data set,are used for experimental performance evaluation to prove the superiority of the proposed method.And to prove that high performance improvement can be achieved even with a small training dataset,the ratio of the training dataset was divided into 70%,50%,and 30%for comparative analysis.As a result of applying the mixed contour data augmentation technique,it was possible to achieve a classification accuracy improvement of up to 4.64%and high accuracy even with a small amount of data set.In addition,it is expected that the mixed contour data augmentation technique can be applied in various fields by proving the excellence of the proposed data augmentation technique using benchmark datasets.

Real-Time Digital Image Stabilization Based on Contour

A new digital image stabilization method is proposed for real-time application based on image contour. The image intensities are projected to several gray levels by thresholding before extracting contour points. Matching position could be found using these contour points. All pixels are still used for refined matching near the matching position. This algorithm is more robust against changes in illumination and noise affection. The adaptive global motion judgement can remove the affection of intruding object. All those are realized on normally available PC.

A Deep Learning Hierarchical Ensemble for Remote Sensing Image Classification

Artificial intelligence,which has recently emerged with the rapid development of information technology,is drawing attention as a tool for solving various problems demanded by society and industry.In particular,convolutional neural networks(CNNs),a type of deep learning technology,are highlighted in computer vision fields,such as image classification and recognition and object tracking.Training these CNN models requires a large amount of data,and a lack of data can lead to performance degradation problems due to overfitting.As CNN architecture development and optimization studies become active,ensemble techniques have emerged to perform image classification by combining features extracted from multiple CNN models.In this study,data augmentation and contour image extraction were performed to overcome the data shortage problem.In addition,we propose a hierarchical ensemble technique to achieve high image classification accuracy,even if trained from a small amount of data.First,we trained the UCMerced land use dataset and the contour images for each image on pretrained VGGNet,GoogLeNet,ResNet,DenseNet,and EfficientNet.We then apply a hierarchical ensemble technique to the number of cases in which each model can be deployed.These experiments were performed in cases where the proportion of training datasets was 30%,50%,and 70%,resulting in a performance improvement of up to 4.68%compared to the average accuracy of the entire model.

Non-contact Acoustic Contour Imaging Method Using Phase Difference in Extremely Shallow Underground

Buried object which exists at extremely shallow underground can be detected by using an acoustic vibration and a SLDV （scanning laser Doppler vibrometer）. Non-contact acoustic exploration can be realized by using air coupled sound. It was confirmed that discovery and the identification of a buried thing were possible using the OFR method （optimum frequency range method）. However, in this method, only the amplitude of the vibration speed spectrum is used and did not use the phase information. Therefore a new imaging method is proposed that used the phase information of the vibration speed spectrum. From the experimental results, the possibility of the outline extraction by using the phase difference between each scan point is confirmed. As a future task, the phase difference image will become clearer by changing the scan point density and long duration time of emitted wave.

Formal Photograph Compression Algorithm Based on Object Segmentation

Small storage space for photographs in formal documents is increasingly necessary in today＇s needs for huge amounts of data communication and storage. Traditional compression algorithms do not sufficiently utilize the distinctness of formal photographs. That is, the object is an image of the human head, and the background is in unicolor. Therefore, the compression is of low efficiency and the image after compression is still space-consuming. This paper presents an image compression algorithm based on object segmentation for practical high-efficiency applications. To achieve high coding efficiency, shape-adaptive discrete wavelet transforms are used to transformation arbitrarily shaped objects. The areas of the human head and its background are compressed separately to reduce the coding redundancy of the background. Two methods, lossless image contour coding based on differential chain, and modified set partitioning in hierarchical trees （SPIHT） algorithm of arbitrary shape, are discussed in detail. The results of experiments show that when bit per pixel （bpp）is equal to 0.078, peak signal-to-noise ratio （PSNR） of reconstructed photograph will exceed the standard of SPIHT by nearly 4dB.

Residual stress determination in friction stir butt welded joints using a digital image correlation-aided slitting technique

This paper presents an incremental cutting method for evaluating the longitudinal residual stresses in a butt welded thin plate via combining the traditional residual stress measurement methods and the advanced optical technique.The proposed approach,which can be called digital image correlation（DIC）-aided slitting technique,introduces a successive extension slot to a specimen and employs the DIC technique to measure the released displacement profiles of the cutting sections after each cutting increment.Then the displacement profiles are used to directly calculate the residual stress distributions up to the slot tip and hence,a stress distribution can be obtained after a cutting increment.Finally,all of the stress distributions are averaged to ultimately determine the original residual stress field.This method does not include any complex experimental operations or tedious derivation,and the resolution of stress variation is greatly improved by the continuous measurement of the released displacements.The presented method has been preliminarily verified by a specimen with residual stress introduced by a four-point bending test.The results show that residual stresses determined by the DIC-aided slitting technique agree well with those from finite element（FE） prediction.The residual stress in a friction stir welded aluminum specimen obtained by the presented technique is also consistent with the evaluations given by X-ray diffraction.Furthermore,the residual stresses obtained by the DIC-aided slitting technique demonstrate higher accuracy and stability than the evaluations derived by the DIC-aided contour method.