http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Transfer Learning 기법을 이용한 가스 누출 영역 분할 성능 비교
( Marshall ),( Jang-sik Park ),( Seong-mi Park ) 한국산업융합학회 2020 한국산업융합학회 논문집 Vol.23 No.3
Safety and security during the handling of hazardous materials is a great concern for anyone in the field. One driving point in the security field is the ability to detect the source of the danger and take action against it as quickly as possible. Via the usage of a fully convolutional network, it is possible to create the label map of an input image, indicating what object is occupying the specific area of the image. This research employs the usage of U-net, which was constructed in biomedical field segmentation to segment cells, instead of the original FCN. One of the challenges that this research faces is the availability of ground truth with precise labeling for the dataset. Testing the network after training resulted in some images where the network pronounces even better detail than the expected label map. With better detailed label map, the network might be able to produce better segmentation is something to be studied in further research.
Performance Comparison of Gas Leak Region Segmentation Based on Transfer Learning
Marshall, Marshall,Park, Jang-Sik,Park, Seong-Mi The Korean Society of Industry Convergence 2020 한국산업융합학회 논문집 Vol.23 No.3
Safety and security during the handling of hazardous materials is a great concern for anyone in the field. One driving point in the security field is the ability to detect the source of the danger and take action against it as quickly as possible. Via the usage of a fully convolutional network, it is possible to create the label map of an input image, indicating what object is occupying the specific area of the image. This research employs the usage of U-net, which was constructed in biomedical field segmentation to segment cells, instead of the original FCN. One of the challenges that this research faces is the availability of ground truth with precise labeling for the dataset. Testing the network after training resulted in some images where the network pronounces even better detail than the expected label map. With better detailed label map, the network might be able to produce better segmentation is something to be studied in further research.
PRIMARY MICROBENTHIC ALGAL PRODUCTION IN CHESAPEAKE BAY
MARSHALL, Harold,WENDKER, Susanne,NESIUS, Kneeland 영남대학교 해양과학연구소 1997 Marine Nature Vol.5 No.1
ABSTRACT :The primary production of the microbenthic algae was determined over a 12 month period at 5 mud flat stations in the southern Chesapeake Bay. Productivity rates varied monthly and spatially in the Bay, with highest productivity occurring during summer, and a diverse assemblage of pennate diatoms the dominant algae component.The annual productivity rate of the microbenthic algae was 142.4 gCm-²yr-¹
Marshall, Ryan S.,Chai, Kil-Byoung,Bellan, Paul M. American Astronomical Society 2017 The Astrophysical Journal Vol.837 No.1
<P>The grain growth process in the Caltech water-ice dusty plasma experiment has been studied using a high-speed camera and a long-distance microscope lens. It is observed that (i) the ice grain number density decreases fourfold as the average grain major axis increases from 20 to 80 mu m, (ii) the major axis length has a log-normal distribution rather than a power-law dependence, and (iii) no collisions between ice grains are apparent. The grains have a large negative charge resulting in strong mutual repulsion and this, combined with the fractal character of the ice grains, prevents them from agglomerating. In order for the grain kinetic energy to be sufficiently small to prevent collisions between ice grains, the volumetric packing factor (i.e., ratio of the actual volume to the volume of a circumscribing ellipsoid) of the ice grains must be less than similar to 0.1 depending on the exact relative velocity of the grains in question. Thus, it is concluded that direct accretion of water molecules is very likely to dominate the observed ice grain growth.</P>
ADVANCED TEST REACTOR TESTING EXPERIENCE - PAST, PRESENT AND FUTURE
Marshall Frances M. Korean Nuclear Society 2006 Nuclear Engineering and Technology Vol.38 No.5
The Advanced Test Reactor (ATR), at the Idaho National Laboratory (INL), is one of the world's premier test reactors for providing the capability for studying the effects of intense neutron and gamma radiation on reactor materials and fuels. The physical configuration of the ATR, a 4-leaf clover shape, allows the reactor to be operated at different power levels in the comer 'lobes' to allow for different testing conditions for multiple simultaneous experiments. The combination of high flux (maximum thermal neutron fluxes of 1E15 neutrons per square centimeter per second and maximum fast [E>1.0 MeV] neutron fluxes of 5E14 neutrons per square centimeter per second) and large test volumes (up to 122 cm long and 12.7 cm diameter) provide unique testing opportunities. The current experiments in the ATR are for a variety of test sponsors - US government, foreign governments, private researchers, and commercial companies needing neutron irradiation services. There are three basic types of test configurations in the ATR. The simplest configuration is the sealed static capsule, which places the capsule in direct contact with the primary coolant. The next level of experiment complexity is an instrumented lead experiment, which allows for active control of experiment conditions during the irradiation. The most complex experiment is the pressurized water loop, in which the test sample can be subjected to the exact environment of a pressurized water reactor. For future research, some ATR modifications and enhancements are currently planned. This paper provides more details on some of the ATR capabilities, key design features, experiments, and future plans.