• Journal of information and communication convergence engineering
• /
• v.16 no.3
• /
• pp.173-178
• /
• 2018

Accurate classification of cloud images is a challenging task. Almost all the existing methods rely on hand-crafted feature extraction. Their limitation is low discriminative power. In the recent years, deep learning with convolution neural networks (CNNs), which can auto extract features, has achieved promising results in many computer vision and image understanding fields. However, deep learning approaches usually need large datasets. This paper proposes a deep learning approach for classification of cloud image patches on small datasets. First, we design a suitable deep learning model for small datasets using a CNN, and then we apply data augmentation and dropout regularization techniques to increase the generalization of the model. The experiments for the proposed approach were performed on SWIMCAT small dataset with k-fold cross-validation. The experimental results demonstrated perfect classification accuracy for most classes on every fold, and confirmed both the high accuracy and the robustness of the proposed model.

• Geophysics and Geophysical Exploration
• /
• v.6 no.4
• /
• pp.187-194
• /
• 2003

A small-loop electromagnetic (EM) system using multiple frequencies has advantages in survey speed and cost despite of limitation on its depth of investigation. Therefore, small-loop EM surveys have been frequently used on various site investigations involving engineering and environmental problems. We have developed a subsurface imaging technique using small loop EM data. We used a one-dimensional (ID) inversion method to reconstruct a subsurface image from frequency EM sounding data. Tests using simulated data show that the method can reasonably recover the subsurface resistivity structure. Also, the method was tested on field data obtained with multiple frequency small loop EM system at a farm in Chunchon, Korea. The resistivity image obtained form field data compares favorably with the image from the dipole-dipole resistivity survey.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.180-194
• /
• 2021

For small image registration, feature-based approaches are likely to fail as feature detectors cannot detect enough feature points from low-resolution images. The classic FFT approach's prediction accuracy is high, but the registration time can be relatively long, about several seconds to register one image pair. To achieve real-time and high-precision rigid registration for small images, we apply deep neural networks for supervised rigid transformation prediction, which directly predicts the transformation parameters. We train deep registration models with rigidly transformed CIFAR-10 images and STL-10 images, and evaluate the generalization ability of deep registration models with transformed CIFAR-10 images, STL-10 images, and randomly generated images. Experimental results show that the deep registration models we propose can achieve comparable accuracy to the classic FFT approach for small CIFAR-10 images (32×32) and our LSTM registration model takes less than 1ms to register one pair of images. For moderate size STL-10 images (96×96), FFT significantly outperforms deep registration models in terms of accuracy but is also considerably slower. Our results suggest that deep registration models have competitive advantages over conventional approaches, at least for small images.

• Proceedings of the ESK Conference
• /
• 1998.04a
• /
• pp.87-92
• /
• 1998

It is needed that the small, cheap and low power dissipated image stabilizing system is used in a small camcorder. It is essential to detect the hand shake of human exactly for the good design of image stabilizing system. In this study, the hand shake is measured by the experiment on 17 peoples in peak-to- peak voltage and frequency. The shake of image by hand by hand is independent of human age or weight, and dependent of calmness of character or concentration of mind.

• Proceedings of the IEEK Conference
• /
• 1998.06a
• /
• pp.661-664
• /
• 1998

Since the region segmentation at high resolution contains most of viable semantic object contours in an image, the bottom-up approach for image segmentation is appropriate for the application such as MPEG-4 which needs to preserve semantic object contours. However, the conventioal region merging methods, that follow the region segmentation, have poor performance in keeping low-contrast semantic object contours. In this paper, we propose an image segmentation algorithm based on classified region merging. The algorithm pre-segments an image with a large number of small regions, and also classifies it into several classes having similar gradient characteristics. Then regions only in the same class are merged according to the boundary weakness or statisticsal similarity. The simulation result shows that the proposed image segmentation preserves semantic object contours very well even with a small number of regions.

• Proceedings of the KIEE Conference
• /
• 1995.11a
• /
• pp.225-227
• /
• 1995

An object detection algorithm using a modified IDM(Image Differential Method) is proposed for detecting an object in a level crossing area. The conventional object detection method using LASER light has the deadzone that it cannot detect small objects, while the object detection method using image data in a level crossing area can detect such small objects. But the image data in a level crossing area can be changeable easily because the data is outdoor and sensitive to such surrounding environments as the change of the sun beam, the shadow of cars, and so on. So we resolve these problems by adding the normalization and the process for shadow of the image data in a level crossing area to the basic IDM(Image Differential Method).

• 한국정보디스플레이학회:학술대회논문집
• /
• 2009.10a
• /
• pp.972-974
• /
• 2009

This paper presents smart regional contrast enhancement technique of partitioned image for local dimming backlight on small-sized mobile display to reach two goals. One is to save the power consumption, and the other to improve contrast ratio of display image. Recently new advanced method is proposed, named local dimming method, that backlight LED is positioned on backside of the display panel. So it is important to partition an image by sub blocks and then post-processing independantly. This means regional contrast enhancement. After partitioning, we compare the mean luminance(Y) value of each sub-block image with the one of original whole image. If some blocks have the mean value lower than the one of whole image, they are processed with the proposed method and others are bypassed. Simultaneously the information of the processed blocks are transferred to BLC(Backlight LED Controller). And then the supply current of each backlight LED is reduced to realize the contrast ratio enhancement and at the same time to power consumption reduction. In addition, we verify this proposed method is free from blocking artifacts.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.1
• /
• pp.291-299
• /
• 2021

In 2019, SUV sales surpassed sedans in the domestic sales market with phenomenal domestic sales. The strength of SUVs around the world is expected to continue in the future. South Korea's K-company aggressively launched small SUVs in the SUV market. Its simple lineup is recognized as a brand image, not as a SUV. It is time to evaluate this. Therefore, it influences the purchasing decisions of potential customers and buyers of small SUVs through the evaluation of design images of small SUVs in Korea. Rather than the functional properties of the SUV model, it is purchased by emotional characteristics, brand symbolism, and image. Subconsciousness of the purchasing psychology of the end consumer was used by metaphor extraction techniques. Customers wanted to study the evaluation of small SUV design images that fit their needs. We wanted to see if consumers who intend to purchase or purchase small SUVs in Korea had a connection with the image of design of small SUVs in Korea. The conclusion of the study was extracted through ZMET, a metaphor extraction technique, with the latent consciousness of the primary ambiguous message from the consumer's feeling and representation of the image. Therefore, based on the results of this study, we hope that the images presented in SUVs in the future will be used as a design guide in the development of small SUVs to influence customer thinking and behavior.

• Journal of Korea Multimedia Society
• /
• v.10 no.4
• /
• pp.432-440
• /
• 2007

In this paper, we propose a small target detection algorithm for FLIR image with complex background. First, we compute the motion information of target from the difference between the current frame and the created background image. However, the slow speed of target cause that it has the very low gray level value in the difference image. To improve the gray level value, we perform the local gamma correction for difference image. So, the detection index is computed by using statistical characteristics in the improved image and then we chose the lowest detection index a true target. Experimental results show that the proposed method has significantly the good detection performance.

• Journal of the Institute of Convergence Signal Processing
• /
• v.14 no.1
• /
• pp.33-38
• /
• 2013

In this paper, we propose an algorithm to track multi-small targets in infrared image sequences in case of dissipation or creation of targets by using the background estimation filter, Kahnan filter and mean shift algorithm. We detect target candidates in a still image by subtracting an original image from an background estimation image, and we track multi-targets by using Kahnan filter and target selection. At last, we adjust specific position of targets by using mean shift algorithm In the experiments, we compare the performance of each background estimation filters, and verified that proposed algorithm exhibits better performance compared to classic methods.