• The Journal of Bigdata
• /
• v.7 no.1
• /
• pp.37-47
• /
• 2022

Autonomous delivery operation data is the key to driving a paradigm shift for last-mile delivery in the Corona era. To bridge the technological gap between domestic autonomous delivery robots and overseas technology-leading countries, large-scale data collection and verification that can be used for artificial intelligence training is required as the top priority. Therefore, overseas technology-leading countries are contributing to verification and technological development by opening AI training data in public data that anyone can use. In this paper, 326 objects were collected to trainn autonomous delivery robots, and artificial intelligence models such as Mask r-CNN and Yolo v3 were trained and verified. In addition, the two models were compared based on comparison and the elements required for future autonomous delivery robot research were considered.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.8
• /
• pp.43-51
• /
• 2024

This paper proposes a method for acquiring and analyzing ocular data using a special-purpose diver mask targeted at divers who primarily engage in underwater activities. This involves tracking the user's gaze with the help of a custom-built ocular dataset and a YOLOv8-nano model developed for this purpose. The model achieved an average processing time of 45.52ms per frame and successfully recognized states of eyes being open or closed with 99% accuracy. Based on the analysis of the ocular data, a gaze tracking algorithm was developed that can map to real-world coordinates. The validation of this algorithm showed an average error rate of about 1% on the x-axis and about 6% on the y-axis.

• Journal of Korean Society for Atmospheric Environment
• /
• v.34 no.3
• /
• pp.493-507
• /
• 2018

Due to high spatio-temporal variability of amount and optical/microphysical properties of atmospheric aerosols, satellite-based observations have been demanded for spatiotemporal monitoring the major aerosols. Observations of the heavy aerosol episodes and determination on the dominant aerosol types from a geostationary satellite can provide a chance to prepare in advance for harmful aerosol episodes as it can repeatedly monitor the temporal evolution. A new geostationary observation sensor, namely the Advanced Himawari Imager (AHI), onboard the Himawari-8 platform, has been observing high spatial and temporal images at sixteen wavelengths from 2016. Using observed spectral visible reflectance and infrared brightness temperature (BT), the algorithm to find major aerosol type such as volcanic ash (VA), desert dust (DD), polluted aerosol (PA), and clean aerosol (CA), was developed. RGB color composite image shows dusty, hazy, and cloudy area then it can be applied for comparing aerosol detection product (ADP). The CALIPSO level 2 vertical feature mask (VFM) data and MODIS level 2 aerosol product are used to be compared with the Himawari-8/AHI ADP. The VFM products can deliver nearly coincident dataset, but not many match-ups can be returned due to presence of clouds and very narrow swath. From the case study, the percent correct (PC) values acquired from this comparisons are 0.76 for DD, 0.99 for PA, 0.87 for CA, respectively. The MODIS L2 Aerosol products can deliver nearly coincident dataset with many collocated locations over ocean and land. Increased accuracy values were acquired in Asian region as POD=0.96 over land and 0.69 over ocean, which were comparable to full disc region as POD=0.93 over land and 0.48 over ocean. The Himawari-8/AHI ADP algorithm is going to be improved continuously as well as the validation efforts will be processed by comparing the larger number of collocation data with another satellite or ground based observation data.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.6
• /
• pp.121-126
• /
• 2004

An efficient multi face detection technique is proposed based on hierarchical context-free attention operators in which multiple faces are efficiently detected from a noisy and complex background. A noise-tolerant generalized symmetry transform (NTSGT) is applied hierarchically, as a context free attention operator, to the input pyramidal image for the high speed global location of the regions of face candidates (ROFCs) with a single mask. For the face verification, local NTGST is applied within each ROFC to confirm the existence of the detailed facial features. First, by globally applying NTGST which introduces the average pyramid method and focusing to the input image with complex background, ROFCs with recognizable resolution are detected robustly. Morphological operations are applied only to the each detected ROFCs to emphasize the facial features like eyes and lips. Then, eyes are detected by locally appling NTGST to the ROFCs and only faces are detected by verifying the existence of the geometrical features of the faces relatively to the location of eyes. The experimental results show that the proposed method can efficiently detect multiple faces from a noisy or complex background with 93.5% detection rate.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.8
• /
• pp.106-114
• /
• 2015

Nowadays, cameras embedded in smartphones can take high resolution photographs that can be used to analyze skin conditions without using specialized equipments. In shooting photographs with a smartphone, it is difficult to maintain a uniform shooting distance. Therefore, it is essential to adapt a skin analysis method to the shooting distance. In this paper, we focus on a pore detection algorithm that is adaptive to the camera distance. We develop a relationship model between the shooting distance and the appropriate size of the pore detection mask. In addition, we propose a method to estimate the normalized pore size (i. e. pore size at a standard shooting distance). We conducted experiments on skin images taken from different shooting distances. It was verified that the proposed method can achieve more accurate pore detection result, close to those from skin images taken at a standard shooting distance.

• Journal of Korea Multimedia Society
• /
• v.13 no.3
• /
• pp.457-466
• /
• 2010

In this paper, we propose a new method for detecting digital forgery that identify interpolated region between digital composited images. For detecting the interpolation factor and the tampered regions, we perform two algorithms: The first algorithm is to estimate the interpolation factors using the differential equation for forgery image along the horizontal, vertical, and diagonal directions, respectively; The second algorithm is to scan the interpolation factors along each direction for detection areas as the mask of the optical window size($64{\times}64$) in order to find out the forgery region. A detection map of the forgery is classified with the magnitude of estimated interpolation factors into colors. This detection map can be used to find out interpolated regions from the tampered image. Experimental results demonstrate the proposed algorithms are proven on several examples. We also show the proposed approach is to accurately detect interpolated regions from digital composite images.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2012.02a
• /
• pp.436-437
• /
• 2012

Double patterning technology (DPT) has been suggested as a promising candidates of the next generation lithography technology in FLASH and DRAM manufacturing in sub-40nm technology node. DPT enables to overcome the physical limitation of optical lithography, and it is expected to be continued as long as e-beam lithography takes place in manufacturing. Several different processes for DPT are currently available in practice, and they are litho-litho-etch (LLE), litho-etch-litho-etch (LELE), litho-freeze-litho-etch (LFLE), and self-aligned double patterning (SADP) [1]. The self-aligned approach is regarded as more suitable for mass production, but it requires precise control of sidewall space etch profile for the exact definition of hard mask layer. In this paper, we propose etch end point detection (EPD) in spacer etching to precisely control sidewall profile in SADP. Conventional etch EPD notify the end point after or on-set of a layer being etched is removed, but the EPD in spacer etch should land-off exactly after surface removal while the spacer is still remained. Precise control of real-time in-situ EPD may help to control the size of spacer to realize desired pattern geometry. To demonstrate the capability of spacer-etch EPD, we fabricated metal line structure on silicon dioxide layer and spacer deposition layer with silicon nitride. While blanket etch of the spacer layer takes place in inductively coupled plasma-reactive ion etching (ICP-RIE), in-situ monitoring of plasma chemistry is performed using optical emission spectroscopy (OES), and the acquired data is stored in a local computer. Through offline analysis of the acquired OES data with respect to etch gas and by-product chemistry, a representative EPD time traces signal is derived. We found that the SE-EPD is useful for precise control of spacer etching in DPT, and we are continuously developing real-time SE-EPD methodology employing cumulative sum (CUSUM) control chart [2].

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.55-58
• /
• 2024

In this paper, we propose a design and implementation of big data-based fine dust anomaly detection machine learning system. The proposed is system that classifies the fine dust air quality index through meteorological information composed of fine dust and big data. This system classifies fine dust through the design of an anomaly detection algorithm according to the outliers for each air quality index classification categories based on machine learning. Depth data of the image collected from the camera collects images according to the level of fine dust, and then creates a fine dust visibility mask. And, with a learning-based fingerprinting technique through a mono depth estimation algorithm, the fine dust level is derived by inferring the visibility distance of fine dust collected from the monoscope camera. For experimentation and analysis of this method, after creating learning data by matching the fine dust level data and CCTV image data by region and time, a model is created and tested in a real environment.

• Journal of Advanced Information Technology and Convergence
• /
• v.9 no.1
• /
• pp.115-125
• /
• 2019

In this paper, the image is received from the camera and the lane is sensed. There are various ways to detect lanes. Generally, the method of detecting edges uses a lot of the Sobel edge detection and the Canny edge detection. The minimum use of multiplication and division is used when designing for the hardware configuration. The images are tested using a black box image mounted on the vehicle. Because the top of the image of the used the black box is mostly background, the calculation process is excluded. Also, to speed up, YCbCr is calculated from the image and only the data for the desired color, white and yellow lane, is obtained to detect the lane. The median filter is used to remove noise from images. Intermediate filters excel at noise rejection, but they generally take a long time to compare all values. In this paper, by using addition, the time can be shortened by obtaining and using the result value of the median filter. In case of the Sobel edge detection, the speed is faster and noise sensitive compared to the Canny edge detection. These shortcomings are constructed using complementary algorithms. It also organizes and processes data into parallel processing pipelines. To reduce the size of memory, the system does not use memory to store all data at each step, but stores it using four line buffers. Three line buffers perform mask operations, and one line buffer stores new data at the same time as the operation. Through this work, memory can use six times faster the processing speed and about 33% greater quantity than other methods presented in this paper. The target operating frequency is designed so that the system operates at 50MHz. It is possible to use 2157fps for the images of 640by360 size based on the target operating frequency, 540fps for the HD images and 240fps for the Full HD images, which can be used for most images with 30fps as well as 60fps for the images with 60fps. The maximum operating frequency can be used for larger amounts of the frame processing.

• Journal of the Korea Convergence Society
• /
• v.8 no.11
• /
• pp.159-165
• /
• 2017

We were investigated the content of 7 preservatives for sheet mask samples(n=42) sold in markets of Daejeon metropolitan city in 2016. &3.3%(n=35) of all samples contained at least one of preservatives. In samples of 30.95(n=14) and 2.39%(n=1) was detected with 2 and 3 preservatives. Phenoxyethaol(PE), methylparaben(MP), chlorphenesin(CP) and benzyl alcohol(BA) showed detection rate of 76.19(n=32), 16.67(n=9), 21.43(n=7) and 2.38%(n=1), respectively. Also The content of detected preservative showed range of 0.06~0.71, 0.18~0.35, 0.06~0.71 and 0.32% and was within the maximum allowed amount established by Korean FDA. However ethylparaben(EP), propylparaben(PP) and benzylparaben(BP) in all samples was not detected. These results can be useful for sharing in emotion-fusioned information and supplying right to know of user.