• 천문학회보
• /
• 제41권1호
• /
• pp.42.3-43
• /
• 2016

We present multiband photometric analyses of 10 core-collapse supernovae in the near-infrared and visible wavebands. Our infrared data is from observations of the supernovae using the Wide Field Infrared Camera at the Palomar 5-m telescope as part of the Caltech Core-Collapse Supernova Program, while we obtain the visible data from publicly available data base. By fitting the broadband spectral energy distribution with a black body and, when necessary, modified black body component, we estimate physical parameters of the supernovae more accurately and also conduct a systematic investigation of when the supernovae show any indication of dust formation.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2019년도 하계학술대회
• /
• pp.1-4
• /
• 2019

최근에 딥러닝 기술을 적용한 보행자 검출 연구가 활발히 진행되고 있다. 연구자들은 딥러닝 네트워크를 이용하여 보행자 오검출율을 낮추는 방법에 대해 지속적으로 연구하여 성능을 꾸준히 상승시켰다. 그러나 대부분의 연구는 다중 스케일 보행자가 분포되는 저해상도 영상에서 보행자를 제대로 검출하지 못하는 어려움이 존재한다. 따라서 본 연구에서는 기존의 Faster R-CNN구조를 기반으로 하여 새로운 다중 특징 융합 레이어와 다중 스케일 앵커 박스를 적용하여 보행자 오검출율을 줄이는 MS-FRCNN(Multi-scaleFaster R-CNN)구조를 제안한다. 제안된 방식의 성능 검증을 위해 Caltech 데이터세트를 이용하여 실험한 결과, 제안된 MS-FRCNN방식이 기존의 다른 보행자 검출 방식보다 다중 스케일 보행자 검출에서 medium 조건하에 5%, all 조건하에 3.9% 나아짐을 알 수 있었다.

• 한국광학회:학술대회논문집
• /
• 한국광학회 2000년도 하계학술발표회
• /
• pp.96-98
• /
• 2000

Room-temperature continuous operation of two-dimensional photonic crystal lasers is achieved at 1.6 ${\mu}{\textrm}{m}$ by using InGaAsP slab-waveguide triangular photonic crystal on top of wet-oxidized aluminum oxide. The main difficulty in the realization of photonic bandgap (OBG) structures has been the nontrivial difficulties in nanofabrication, especially for 3-dimensional PBG structures. Recently, 2-D PBG structures have attracted a great deal of attention due to their simplicity in fabrication and theoretical study as compared to the three-dimensional counterparts [1]. Recently, air-gulfed 2-D slab PBG lasers were reported by Caltech group [2]. However, this air-slab structure is mechanically fragile and thermally unforgiving. Therefore, a new structure that can remove this thermal limitation is dearly sought after for 2-D PBG laser to have practical meaning. In this talk, we report room-temperature continuous operation of 2-D photonic bandgap lasers that are thermally and mechanically stable.

• 천문학회보
• /
• 제45권1호
• /
• pp.39.2-39.2
• /
• 2020

Based upon the previous heritage in the complete development of the infrared imaging spectrometer, NISS (Near-infrared Imaging Spectrometer for Star formation history) onboard NEXTSat-1, we are participating in the NASA MIDEX mission (PI Institute: Caltech), the all-sky infrared spectro-photometric surveyor SPHEREx (Spectro-Photometer for the History of the Universe, Epoch of Reionization, and Ices Explorer). The SPHEREx will provide us the first all-sky infrared spectro-photometric data set to probe the origin of our Universe, to explore the origin and evolution of galaxies, and to explore whether planets around other stars could harbor life. After the SPEHREx have passed the PDR (Preliminary Design Review) on this September, the fabrication of flight hardware will be started soon. As an international partner, KASI takes part in the hardware development, the operation and the science for the SPHEREx. Here, we report the status of the SPHEREx project and the progress in the Korean participation.

• Journal of Electrical Engineering and Technology
• /
• 제13권6호
• /
• pp.2468-2478
• /
• 2018

We propose the Sparse Feature Convolutional Neural Network (SFCNN) to reduce the volume of convolutional neural networks (CNNs). Despite the superior classification performance of CNNs, their enormous network volume requires high computational cost and long processing time, making real-time applications such as online-training difficult. We propose an advanced network that reduces the volume of conventional CNNs by producing a region-based sparse feature map. To produce the sparse feature map, two complementary region-based value extraction methods, cluster max extraction and local value extraction, are proposed. Cluster max is selected as the main function based on experimental results. To evaluate SFCNN, we conduct an experiment with two conventional CNNs. The network trains 59 times faster and tests 81 times faster than the VGG network, with a 1.2% loss of accuracy in multi-class classification using the Caltech101 dataset. In vehicle classification using the GTI Vehicle Image Database, the network trains 88 times faster and tests 94 times faster than the conventional CNNs, with a 0.1% loss of accuracy.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권1호
• /
• pp.364-380
• /
• 2016

The most popular approach in object classification is based on the bag of visual-words model, which has several fundamental problems that restricting the performance of this method, such as low time efficiency, the synonym and polysemy of visual words, and the lack of spatial information between visual words. In view of this, an object classification based on weakly supervised E2LSH and saliency map weighting is proposed. Firstly, E2LSH (Exact Euclidean Locality Sensitive Hashing) is employed to generate a group of weakly randomized visual dictionary by clustering SIFT features of the training dataset, and the selecting process of hash functions is effectively supervised inspired by the random forest ideas to reduce the randomcity of E2LSH. Secondly, graph-based visual saliency (GBVS) algorithm is applied to detect the saliency map of different images and weight the visual words according to the saliency prior. Finally, saliency map weighted visual language model is carried out to accomplish object classification. Experimental results datasets of Pascal 2007 and Caltech-256 indicate that the distinguishability of objects is effectively improved and our method is superior to the state-of-the-art object classification methods.

• 천문학회보
• /
• 제41권2호
• /
• pp.57.1-57.1
• /
• 2016

Methanol ($CH_3OH$) is a key species in the formation of complex organic molecules. We report the first detection of solid $CH_3OH$ in a line of sight toward the Galactic center (GC) region, based on L-band spectra taken with the Subaru telescope, aided by L'-band imaging data and moderate-resolution spectra from NASA/IRTF. It is found toward a background star, ~8000 AU in projected distance from a newly discovered massive young stellar object (YSO). This YSO also exhibits a strong $CO_2$ ice absorption band at ${\sim}15{\mu}m$ in Spitzer/IRS data, which has a prominent long-wavelength wing. It confirms that a high $CH_3OH$ abundance is responsible for the broad $15{\mu}m$ $CO_2$ ice absorption towards massive YSOs in the GC. Clearly, $CH_3OH$ formation in ices is efficient in the GC region, as it is in star-forming regions in the Galactic disk. We discuss implications of our result on the astrochemical processes in the hostile GC molecular clouds.

• 천문학회보
• /
• 제37권1호
• /
• pp.64.1-64.1
• /
• 2012

Cosmic Infrared Background Experiment (CIBER)는 적외선 카메라 및 분광기를 NASA Sounding Rocket에 탑재, 발사하여 적외선우주배경복사를 관측하는 과제이다. CIBER1은 2006년 NASA의 공식 과제로 승인되어, 미국의 Caltech, 한국의 KASI, 일본의 ISAS/JAXA가 국제협력으로 진행되었으며, 2009년 2월 25일, 2010년 7월 10일, 그리고 2012년 2월 25일에 미국 화이트샌드 미사일 기지에서 NASA 사운딩 로켓에 의해 성공적으로 발사되어 우주관측에 성공하였다. CIBER2는 CIBER1 보다 약 10 배 이상의 성능을 가지는 적외선카메라로써 한국의 KASI는 CIBER2 개발에서 광학계 및 광기계부 개발, 전자부 개발에 참여하고 있다. CIBER2는 2012년에 개발을 시작하여 2013년과 2014년에 각각 발사될 예정이다.

• 천문학회보
• /
• 제42권1호
• /
• pp.55.3-55.3
• /
• 2017

The L1641 cloud in Orion is an active site of star formation. We mapped a square region of 60 arcmin by 60 arcmin in the continuum emission from 0.89 mm to 2.0 mm wavelength using MUSIC mounted on the Caltech Submillimeter Observatory 10.4 m telescope. Eight sources were detected in at least two wavelength bands, and all the detected emission comes from thermal dust continuum radiation of dense cloud cores. Their spectral energy distributions were characterized. The dust emissivity spectral index is beta = 1.3 on average, within the range of typical cores in nearby star-forming regions. Two cores, V380 Ori NE and HH 34 MMS, have unusually low emissivity index of beta = 0.3. These cores may contain millimeter-sized dust grains, which suggests that the lifetime of some dense cores can be much longer than the free-fall timescale.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권3호
• /
• pp.522-538
• /
• 2013

Content based image retrieval has become an increasingly important research topic for its wide application. It is highly challenging when facing to large-scale database with large variance. The retrieval systems rely on a key component, the predefined or learned similarity measures over images. We note that, the similarity measures can be potential improved if the data distribution information is exploited using a more sophisticated way. In this paper, we propose a similarity measure learning approach for image retrieval. The similarity measure, so called Fisher kernel, is derived from the probabilistic distribution of images and is the function over observed data, hidden variable and model parameters, where the hidden variables encode high level information which are powerful in discrimination and are failed to be exploited in previous methods. We further propose a discriminative learning method for the similarity measure, i.e., encouraging the learned similarity to take a large value for a pair of images with the same label and to take a small value for a pair of images with distinct labels. The learned similarity measure, fully exploiting the data distribution, is well adapted to dataset and would improve the retrieval system. We evaluate the proposed method on Corel-1000, Corel5k, Caltech101 and MIRFlickr 25,000 databases. The results show the competitive performance of the proposed method.