Hybrid negative example selection using visual and conceptual features

An application of Query-By-Example (QBE) is presented where shots that are visually similar to provided example shots are retrieved. To implement QBE, counter-example shots are required to accurately distinguish shots that are relevant to the query from those that are not (Li and Snoek (2009), Yu et al. (2004)). However, there are usually a huge number of shots, not relevant to a particular query, which can serve as counter-example shots. It is difficult for a user to provide counter-example shots that would aid retrieval. Thus, we developed a QBE method based on partially supervised learning where a retrieval model is constructed by selecting counter-example shots from shots without user supervision. To ensure the speed and accuracy of the QBE method, we select a small number of counter-example shots that are visually similar to given example shots but irrelevant to the query. Such shots are useful for characterizing the boundary between relevant and irrelevant shots. For our method, we first filter shots that are visually dissimilar to example shots based on SVMs on a visual feature. Then we filter shots relevant to the query based on concept detection results from pre-constructed classifiers. Shots that pass the above two tests are considered as counter-example shots. Experimental results obtained using TRECVID 2009 video data validate the effectiveness of our method.     

Complete Mining of Frequent Patterns from Graphs: Mining Graph Data

Basket Analysis, which is a standard method for data mining, derives frequent itemsets from database. However, its mining ability is limited to transaction data consisting of items. In reality, there are many applications where data are described in a more structural way, e.g. chemical compounds and Web browsing history. There are a few approaches that can discover characteristic patterns from graph-structured data in the field of machine learning. However, almost all of them are not suitable for such applications that require a complete search for all frequent subgraph patterns in the data. In this paper, we propose a novel principle and its algorithm that derive the characteristic patterns which frequently appear in graph-structured data. Our algorithm can derive all frequent induced subgraphs from both directed and undirected graph structured data having loops (including self-loops) with labeled or unlabeled nodes and links. Its performance is evaluated through the applications to Web browsing pattern analysis and chemical carcinogenesis analysis.     

Interface control for resolved CFD-DEM with capillary interactions

In this work, a resolved CFD–DEM coupling model for the simulation of gas-liquid-solid flows is developed: the interface capturing method based on the colour function is employed for fluids (i.e. a gas and liquid) whilst Discrete Element Method (DEM) is used for particles. The Volume Penalisation (VP) method is adopted to consider the hydrodynamic interactions between fluids and particles along with the Immersed Free Surface (IFS) method, which artificially extends the gas-liquid interface into the interior of the particle to account for the wettability. The unique point of the proposed model is that the thickness of the gas-liquid interface can be controlled by using both interface compression and diffuse interface techniques simultaneously. From the simulation results, it is presented that the accurate evaluation of the surface tension force as well as the capillary force can be achieved by appropriately controlling the interface thickness. Moreover, the major two methods in the literature to calculate the capillary force are compared in this work. The validity of the proposed model is presented for both static and dynamic cases. The behaviour of two colliding particles with a dynamic liquid bridge is then simulated to demonstrate the applicability of the proposed model to a complex three-phase system.     

Grain-growth-inhibiting effects of primary inclusion particles of ZrO2 and MgO in Fe-10 mass pct Ni alloy

Grain-growth inhibition in an Fe-10 mass pct Ni alloy, which was continuously cooled from a melt, was studied at 1673 K in the presence of primary deoxidation products of ZrO₂ or MgO particles. The mean grain size and grain-size distribution in a cross section were measured as a function of holding time for up to 240 minutes. The grain growth was strongly inhibited by the inclusion particles and was influenced by the dissolved Zr. In the Zr deoxidation, the number of particles per unit area (N _A) ranged from 80 to 650 mm⁻², the ZrO₂ particle size ( ) varied from 1.1 to 1.6 μm, and the dissolved Zr level was below 1800 mass ppm. In the Mg deoxidation, the particle-number density was 90 to 270 mm⁻², the MgO particle size was 1.1 to 1.7 μm, and the dissolved Mg level was below 20 mass ppm. In a logarithmic plot of the ratio of limiting mean grain diameter ( ) to the mean particle diameter ( ) against the volume fraction of particles (f _V), both the value for a given f _V value, which ranged from 0.014 to 0.074 pct, and the slope were significantly lower than that predicted from the two-dimensional relation =(4/π) · f _V ^/−1 , i.e., Zener’s limit. This discrepancy is discussed in light of the fraction of particles at the grain boundaries measured experimentally. Normal grain growth was confirmed from the grain-size distribution observed as a function of holding time, which was best described by the log-normal distribution.     

Short-term algal toxicity test based on phosphate uptake

In order to develop a short-term algal toxicity test, the growth of and the phosphate uptake by the green alga Selenastrum capricornutum during batch culture were observed. In the control medium, S. capricornutum took up phosphate earlier than it grew. It was also observed that the phosphate uptake was inhibited by the presence of a toxicant. From these results, phosphate uptake was considered as one of the useful effect parameters for a short-term algal toxicity test. As the removal rate of phosphate from the medium is a function of the amount of algal cell initially inoculated, the test period is variable. The relationship between the amount of inoculation and phosphate uptake was examined and the test conditions suitable for a 3-h toxicity test were established as one example. According to this test procedure, the inhibitory effect of some toxicants on the phosphate uptake was determined. For comparison, a conventional algal assay based on algal growth was also performed. The EC50s for both tests were close. This indicated that the algal toxicity test method proposed in this paper would be useful for the uses where rapidity is required.     

Reduction of depth variation in an Si etching process by applying an optimized run-to-run control system

A reduction of depth variation in a shallow trench isolation process is desired for obtaining more stable electrical performances of silicon devices. By applying an optimized run-to-run control system to a silicon trench etching process, trench depth variation of the process was reduced. Cp/Cpk of the trench depth of product wafers were improved from 1.10/0.83 to 1.39/1.34 by the control.