• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.16 no.4
• /
• pp.246-253
• /
• 2016

Infrared (IR) imaging has been researched for various applications such as surveillance. IR radiation has the capability to detect thermal characteristics of objects under low-light conditions. However, automatic segmentation for finding the object of interest would be challenging since the IR detector often provides the low spatial and contrast resolution image without color and texture information. Another hindrance is that the image can be degraded by noise and clutters. This paper proposes multi-level segmentation for extracting regions of interest (ROIs) and objects of interest (OOIs) in the IR scene. Each level of the multi-level segmentation is composed of a k-means clustering algorithm, an expectation-maximization (EM) algorithm, and a decision process. The k-means clustering initializes the parameters of the Gaussian mixture model (GMM), and the EM algorithm estimates those parameters iteratively. During the multi-level segmentation, the area extracted at one level becomes the input to the next level segmentation. Thus, the segmentation is consecutively performed narrowing the area to be processed. The foreground objects are individually extracted from the final ROI windows. In the experiments, the effectiveness of the proposed method is demonstrated using several IR images, in which human subjects are captured at a long distance. The average probability of error is shown to be lower than that obtained from other conventional methods such as Gonzalez, Otsu, k-means, and EM methods.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.6
• /
• pp.493-497
• /
• 2016

Infrared (IR) imaging is capable of detecting targets that are not visible at night, thus it has been widely used for the security and defense system. However, the quality of the IR image is often degraded by low resolution and noise corruption. This paper addresses target segmentation with the IR image. Multiple regions of interest (ROI) are extracted by the multi-level segmentation and targets are segmented from the individual ROI. Each level of the multi-level segmentation is composed of a k-means clustering algorithm an expectation-maximization (EM) algorithm, and a decision process. The k-means clustering algorithm initializes the parameters of the Gaussian mixture model (GMM) and the EM algorithm iteratively estimates those parameters. Each pixel is assigned to one of clusters during the decision. This paper proposes the selection and the merging of the extracted ROIs. ROI regions are selectively merged in order to include the overlapped ROI windows. In the experiments, the proposed method is tested on an IR image capturing two pedestrians at night. The performance is compared with conventional methods showing that the proposed method outperforms others.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.8
• /
• pp.3965-3983
• /
• 2017

We propose a novel eye tracking method that can estimate 3D world coordinates using an infrared (IR) stereo camera for indoor and outdoor environments. This method first detects dark evidences such as eyes, eyebrows and mouths by fast multi-level thresholding. Among these evidences, eye pair evidences are detected by evidential reasoning and geometrical rules. For robust accuracy, two classifiers based on multiple layer perceptron (MLP) using gradient local binary patterns (GLBPs) verify whether the detected evidences are real eye pairs or not. Finally, the 3D world coordinates of detected eyes are calculated by region-based stereo matching. Compared with other eye detection methods, the proposed method can detect the eyes of people wearing sunglasses due to the use of the IR spectrum. Especially, when people are in dark environments such as driving at nighttime, driving in an indoor carpark, or passing through a tunnel, human eyes can be robustly detected because we use active IR illuminators. In the experimental results, it is shown that the proposed method can detect eye pairs with high performance in real-time under variable illumination conditions. Therefore, the proposed method can contribute to human-computer interactions (HCIs) and intelligent transportation systems (ITSs) applications such as gaze tracking, windshield head-up display and drowsiness detection.

• ETRI Journal
• /
• v.46 no.5
• /
• pp.851-864
• /
• 2024

Deep learning (DL) has significantly advanced artificial intelligence (AI); however, frameworks such as PyTorch, ONNX, and TensorFlow are optimized for general-purpose GPUs, leading to inefficiencies on specialized accelerators such as neural processing units (NPUs) and processing-in-memory (PIM) devices. These accelerators are designed to optimize both throughput and energy efficiency but they require more tailored optimizations. To address these limitations, we propose the NEST compiler (NEST-C), a novel DL framework that improves the deployment and performance of models across various AI accelerators. NEST-C leverages profiling-based quantization, dynamic graph partitioning, and multi-level intermediate representation (IR) integration for efficient execution on diverse hardware platforms. Our results show that NEST-C significantly enhances computational efficiency and adaptability across various AI accelerators, achieving higher throughput, lower latency, improved resource utilization, and greater model portability. These benefits contribute to more efficient DL model deployment in modern AI applications.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2008.10a
• /
• pp.1120-1123
• /
• 2008

Efficient, color stable multi-EML WOLED have been fabricated using newly synthesized yellowish green dopant Ir(chpy)3 or Ir(mchpy)3. The devices have high external quantum efficiency of 11.7%, color rendering index of 87, variation of CIE coordinate of (0.02, 0.01) between 10 to 5000 cd/m2, and low roll-off in efficiency with increasing brightness.

• The Journal of Korean Association of Computer Education
• /
• v.8 no.6
• /
• pp.113-121
• /
• 2005

Since documents on the Web are naturally partitioned into many document databases, the efficient information retrieval (IR) process requires identifying the document databases that are most likely to provide relevant documents to the query and then querying the identified document databases. We first introduce a neural net agent for such an efficient IR, and then propose the hierarchically organized multi-agent IR system in order to scale our agent with the large number of document databases. In this system, the hierarchical organization of neural net agents reduced the total training cost at an acceptable level without degrading the IR effectiveness in terms of precision and recall. In the experiment, we introduce two neural net IR systems based on single agent approach and multi-agent approach respectively, and evaluate the performance of those systems by comparing their experimental results to those of the conventional statistical systems.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.11
• /
• pp.2759-2764
• /
• 2014

Smoke detector is commonly used to reduce fire detection time. However, technical problems regarding its inaccuracy of laser beam-receiving point on the surface of the sensor associated with incoming interference are identified when the laser transmitter and receiver are installed at a distance of about 100m. In this paper, we propose the auto focus alignment algorithm with high precision to adjust tilting angle of lasers caused by environmental interference so that solve existing issues using multi-level worm gear set.

• The Bulletin of The Korean Astronomical Society
• /
• v.37 no.2
• /
• pp.74.1-74.1
• /
• 2012

We observed the radio sources found from the First Look Survey (FLS) field at the 1.4 GHz radio continuum emission with the Very Large Array (VLA) using the A configuration. We identify point sources and multi component sources at ${\geq}4{\sigma}$ level. We also present the submillimeter properties of the selected radio sources in the FLS field from the Herschel/SPIRE 250/350/500/${\mu}m$ and AzTEC 1.1mm surveys. The counterparts of the radio sources at submillimeter for these called 'submillimeter galaxies (SMGs)' are detected at infrared wavelength with the Spitzer MIPS 24 & 70 ${\mu}m$ sources. Based on the MMT/HECTOSPEC red-shift survey, IRS spectroscopy, and SDSS photometric red-shift survey, the radio sources are likely to be the extragalactic sources. Here, we use the star formation rate (SFR) derived from the MIPS 24 and 70 ${\mu}m$ luminosity to compare the measured SFR from the VLA 1.4 GHz luminosity. These results show that a tight correlation between the SFR from the radio luminosity and the MIPS $24{\mu}m$ rather than that from the MIPS $70{\mu}m$ luminosity. Radio and IR correlation is also used to indicate the radio and IR properties of star-formation in the galaxies and active galactic nuclei (AGNs). Using the counterpart sources selected at IR and radio wavelengths, we employ the IR/radio flux ratios to determine the properties and population of the selected galaxies.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.44 no.10
• /
• pp.47-54
• /
• 2007

Since impulse-radio-based ultra wideband (IR-UWB) do not use carrier frequency but use very short pulse to transmit data it sends data not continuously but discretely and this feature gives the potential to reduce collision in multi-user environment. In this paper, we analyse characteristic of IR-UWB and propose a new collision paradigm, Collision Distribution which changes collision level from packet to pulse. In Collision Distribution mechanical each node sends data with its own pulse interval in random time, distributed manner. It prevents packet drop due to packet collision. We show that Collision Distribution can reduce packet error and can provide real time packet transmission with analysis.

• Korean Chemical Engineering Research
• /
• v.46 no.1
• /
• pp.118-123
• /
• 2008

In the present study, the effect of oxyfluorination treatment on multi-walled nanotubes (MWNTs) supports was investigated by analyzing surface functional groups. The surface characteristics were determined by Fourier transformed-infrared (FT-IR) and X-ray photoelectron spectroscopy (XPS). After the deposition of platinum nanoparticles on the above treated carbon supports, a crystalline size and a loading level had been investigated. Electrochemical properties of the treated MWNTs-supported Pt (Pt/MWNTs) catalysts were analyzed by current-voltage curve measurements. From the results of surface analysis, an oxygen and fluorine-containing functional group had been introduced to the surface of carbon supports. The oxygen and fluorine contents were the highest value at the treatment of 100 temperature. The Pt/100-MWNTs showed the smallest particle crystalline size of 3.5 nm and the highest loading level of 9.4% at the treatment of 100 temperature. However, the sample treated at the higher temperature showed the larger crystalline size and the lower loading level. This indicated that the crystalline size and the loading level could be controlled by changing the temperature of oxyfluorination treatment. Accordingly, an electrochemical activity was enhanced by increasing the temperature of treatment upto 100, and then decreased in the case of 200 and 300. The highest specific current density of 120 mA/mg had been obtained in the case of Pt/100-MWNTs.