• The Journal of Korean Institute of Next Generation Computing
• /
• v.15 no.5
• /
• pp.53-63
• /
• 2019

To provide a hand gesture interface through deep learning in mobile environments, research on the light-weighting of networks is essential for high recognition rates while at the same time preventing degradation of execution speed. This paper proposes a method of real-time recognition of written characters in the air using a finger on mobile devices through the light-weighting of deep-learning model. Based on the SSD (Single Shot Detector), which is an object detection model that utilizes MobileNet as a feature extractor, it detects index finger and generates a result text image by following fingertip path. Then, the image is sent to the server to recognize the characters based on the learned OCR model. To verify our method, 12 users tested 1,000 words using a GALAXY S10+ and recognized their finger with an average accuracy of 88.6%, indicating that recognized text was printed within 124 ms and could be used in real-time. Results of this research can be used to send simple text messages, memos, and air signatures using a finger in mobile environments.

• Korean Journal of Optics and Photonics
• /
• v.33 no.6
• /
• pp.310-316
• /
• 2022

In this paper, the difference in modulation transfer function (MTF) results according to the change in the angle of a slant target when measuring a presampled MTF was confirmed, and the difference was reduced by fitting the edge spread function graph obtained to reduce the error by the target's rotation. Due to the feature of the presampled MTF method, the spatial frequency changed due to the sensor's projected intensity being changed by the target's rotation, and it was confirmed that the difference in the MTF value occurred depending on the rotation angle of the target. In this paper, the MTF was calculated after fitting only one column of the acquired image. It was confirmed that the rotation error is smaller compared to the case of the presampled MTF method and this fitting method can be applied to a scene that contains various target angles, such as auto-focusing using the MTF.

• Current Optics and Photonics
• /
• v.8 no.4
• /
• pp.406-415
• /
• 2024

The major error factors that degrade the efficiency of coherent beam combining (CBC) are numerically studied in a comprehensive manner, paying particular attention to phase, tip-tilt, polarization angle, and beam quality. The power in the bucket (PIB), normalized to the zero-error PIB, is used as a figure of merit to quantify the effect of each error factor. To maintain a normalized PIB greater than or equal to 95% in a 3-channel CBC configuration, the errors in phase, tip-tilt, and polarization angle should be less than 1.06 radians, 1.25 ㎛, and 1.06 radians respectively, when each of the three parameters is calculated independently with the other two set to zero. In a worst-case scenario of the composite errors within the parameter range for the independent-95%-normalized-PIB condition, the aggregate effect would reduce the normalized PIB to 83.8%. It is noteworthy that the PIB performances of a CBC system, depending on phase and polarization-angle errors, share the same characteristic feature. A statistical approach for each error factor is also introduced, to assess a CBC system with an extended number of channels. The impact of the laser's beam-quality factor M² on the combining efficiency is also analyzed, based on a super-Gaussian beam. When M² increases from 1 to 1.3, the normalized PIB is reduced by 2.6%, 11.8%, 12.8%, and 13.2% for a single-channel configuration and 3-, 7-, and 19-channel CBC configurations respectively. This comprehensive numerical study is expected to pave the way for advances in the evaluation and design of multichannel CBC systems and other related applications.

• Journal of KIISE:Software and Applications
• /
• v.29 no.8
• /
• pp.573-583
• /
• 2002

Pattern recognition process is made up of the feature extraction in the pre-processing, the pattern clustering by training and the recognition process. This paper presents the F-LVQ (Fuzzy Learning Vector Quantization) pattern classification model which is optimized by the fuzzy membership function for the OCR(Optical Character Recognition) system. We trained 220 numeric patterns of 22 Hangul and English fonts and tested 4840 patterns whose forms are changed variously. As a result of this experiment, it is proved that the proposed model is more effective and robust than other typical LVQ models.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.3
• /
• pp.41-46
• /
• 2002

We generally used electric optical sensors in order to detect and identify sea objects efficiently. However due to the limitation of view-angle, the region acquired from the sense is restricted. So it is necessary to generate panoramic image from sea images acquired from pan-tilt camera. Previous mosaicing method is not able to generate panoramic image for sea environment because intensity is similar to all region and time varying. In this paper, we proposed new algorithm for generating high-resolution panoramic image on sea environment. Proposed algorithm use single-view point model, applying mosaicing result in feature environment to in sea environment, we overcame the limitation of previous method. In the result of virtual and real experiment, we showed that proposed algorithm is efficient for generating sea panoramic image.

• Korean Journal of Optics and Photonics
• /
• v.23 no.1
• /
• pp.36-41
• /
• 2012

Total external reflection (TER), which does not occur on a dielectric interface, is a unique feature of surface plasmon-polaritons (SPP). We propose an SPP-TER waveguide structure consisting of low-index dielectric nanocore covered with high-index dielectric on a flat metal surface. The SPP mode confined in the nanocore by the TER effect has a mode size much smaller than wavelength scale. Numerical comparison of mode characteristics between the SPP-TER waveguides and other total-internal-reflection-based waveguides such as metal or high-index dielectric nanowires show that the SPP-TER structures can possess higher modal gain for applications of nanocavity lasers.

• Journal of Sensor Science and Technology
• /
• v.4 no.4
• /
• pp.10-15
• /
• 1995

Organic thin film electroluminescent(EL) cells were fabricated. Their output characteristics and luminance versus voltage characteristics were measured with different work function metal electrodes. The EL structure was Indium-Tin-Oxide(ITO)/hole transport layer/emission layer(electron transport layer)/metal electrode. PMMA+TPD(0.5 wt%), MC homopolymer+TPD(0.005 wt%) and (MC/MMA) copolymer+TPD(0.005 wt%) were used as hole transport layer. Ca, Mg, Mg:Ag(10:l) and Al were used as metal electrode. I-V output showed exponential feature, and the threshold voltage of 5 volts and the luminance of over 700 $Cd/m^{2}$ at 10 volts were observed.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.4
• /
• pp.763-781
• /
• 2011

We present a texture-based hatching technique for color images and video. Whereas existing approaches produce monochrome hatching effects in considering of triangular mesh models by applying strokes of uniform size, our scheme produces color hatching effects from photographs and video using strokes with a range of sizes. We use a Delaunay triangulation to create a mesh of triangles with sizes that reflect the structure of an input image. At each vertex of this triangulation, the flow of the image is analyzed and a hatching texture is then created with the same alignment, based on real pencil strokes. This texture is given a modified version of a color sampled from the image, and then it is used to fill all the triangles adjoining the vertex. The three hatching textures that accumulate in each triangle are averaged and the result of this process across all the triangles forms the output image. We can also add a paper texture effect and enhance feature lines in the image. Our algorithm can also be applied to video. The results are visually pleasing hatching effects similar to those seen in color pencil drawings and oil paintings.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.478-481
• /
• 2005

The PMU (Payload Management Unit) is the main subsystem for the management, control and power supply of the MSC (Multi-Spectral Camera) Payload operation. It is the most important function for the electro-optical camera system that performs the Non-Uniformity Correction (NUC) function of the raw imagery data, rearranges the data from the CCD (Charge Coupled Device) detector and output it to the Data Compression and Storage Unit (DCSU). The NUC board in PMU performs it. In this paper, the NUC board system is described in terms of the configuration and the function, the efficiency for non-uniformity correction, and the influence of the data compression upon the peculiar feature of the CCD pixel. The NUC board is an image-processing unit within the PMU that receives video data from the CEV (Camera Electronic Unit) boards via a hotlinkand performs non-uniformity corrections upon the pixels according to commands received from the SBC (Single Board Computer) in the PMU. The lossy compression in DCSU needs the NUC in on-orbit condition.

• Journal of Mechanical Science and Technology
• /
• v.19 no.3
• /
• pp.887-893
• /
• 2005

The aggregability of red blood cells (RBCs) was determined by laser backscattering light analysis in a microfluidic channel. Available techniques for RBC aggregation often adopt a rotational Couette-flow using a bob-and-cup system for disaggregating RBCs, which causes the system to be complex and expensive. A disposable microfluidic channel and vibration generating mechanism were used in the proposed new detection system for RBC aggregation. Prior to measurement, RBC aggregates in a blood sample were completely disaggregated by the application of vibration-induced shear. With the present apparatus, the aggregation indexes of RBCs can be measured easily with small quantities of a blood sample. The measurements with the present aggregometer were compared with those of LORCA and the results showed a strong correlation between them. The aggregability of the defibrinogenated blood RBCs is markedly lower than that of the normal RBCs. The noble feature of this design is the vibration-induced disaggregation mechanism, which can incorporate the disposable element that holds the blood sample.