Food recognition using neural network classifier and multiple hypotheses image segmentation

ABSTRACT

This paper proposes the multiple-hypotheses image segmentation and feed-forward neural network classifier for food recognition to improve the performance. Initially, the food or meal image is given as input. Then, the segmentation is applied to identify the regions, where a particular food item is located using salient region detection, multi-scale segmentation, and fast rejection. Then, the features of every food item are extracted by the global feature and local feature extraction. After the features are obtained, the classification is performed for each segmented region using a feed-forward neural network model. Finally, the calorie value is computed with the aid of (i) food volume and (ii) calorie and nutrition measure based on mass value. The experimental results and performance evaluation are validated. The outcome of the proposed method attains 0.947 for Macro Average Accuracy (MAA) and 0.959 for Standard Accuracy (SA), which provides better classification performance.     

测井资料在优选压裂层位中的应用

压裂是低孔、低渗油层改造的主要增产措施。根据常规测井资料、岩石物理实验分析数据及试油成果。研究储层四性关系，确定区块的低孔、低渗储层评价标准。进一步对油层进行分类，有针对性地为油层压裂提供施工层位和准确的储层参数。通过商543区块实例，阐述了选择压裂层位的方法。     

接轨站设置位置和布置图型的影响因素及其分类

通过对接轨站设置位置和布置图型主要影响因素的分析。按照能够充分反映接轨站作业特点的交接方式。对接轨站所采用的布置图型进行了归纳和分类。     

Biotransformation of l‐menthol by twelve isolates of soil‐borne plant pathogenic fungi (Rhizoctonia solani) and classification of fungi

The microbial transformation of l‐menthol ( 1 ) was investigated by using 12 isolates of soil‐borne plant pathogenic fungi, Rhizoctonia solani (AG‐1‐IA Rs24, Joichi‐2, RRG97‐1; AG‐1‐IB TR22, R147, 110.4; AG‐1‐IC F‐1, F‐4, P‐1; AG‐1‐ID RCP‐1, RCP‐3, and RCP‐7) as a biocatalyst. Rhizoctonia solani F‐1, F‐4 and P‐1 showed 89.7–99.9% yields of converted product from 1 , RCP‐1, RCP‐3, and RCP‐7 26.0–26.9% and the other isolates 0.1–12.0%. In the cases of F‐1, F‐4 and P‐1, substrate 1 was converted to (?)‐(1S,3R,4S,6S)‐6‐hydroxymenthol ( 2 ), (?)‐(1S,3R,4S)‐1‐hydroxymenthol ( 3 ) and (+)‐(1S,3R,4R,6S)‐6,8‐dihydroxymenthol ( 4 ), which was a new compound. Substrate 1 was converted to 2 and/or 3 by RRG97‐1, 110.4, RCP‐1, RCP‐3 and RCP‐7. The structures of the metabolic products were elucidated on the basis of their spectral data. In addition, metabolic pathways of the biotransformation of 1 by Rhizoctonia solani are discussed. Finally, from the main component analysis and the differences in the yields of converted product from 1 , the 12 isolates of Rhizoctonia solani were divided into three groups based on an analysis of the metabolites. Copyright © 2003 Society of Chemical Industry     

低温脂肪酶产生菌的筛选、鉴定及其部分酶学性质

从南极乔治王岛冻土来源的76株低温细菌中筛选到13株低温脂肪酶产生菌，对其中的BTsl0022菌株进行鉴定。通过生理生化特征、16s rDNA基因序列的同源性和系统发育分析发现，菌株RTsl0022属于假单胞菌属(Pseudomonas)，但与已定名的假单胞菌有一定的差异，与未定名的Pseudomonas sp．PsB的亲缘关系最接近，故将其暂定名为Pseudomonas sp．BTsl0022。对该菌脂肪酶的酶学性质初步研究表明，酶的最适作用温度为24℃，对热敏感，60℃处理30min仅残留25％酶活性，酶的适宜作用pH范围在7．0～9．0，最适pH为8．0。     

An intelligent web-page classifier with fair feature-subset selection

The explosion of on-line information has given rise to many manually constructed topic hierarchies (such as Yahoo!!). But with the current growth rate in the amount of information, manual classification in topic hierarchies results in an immense information bottleneck. Therefore, developing an automatic classifier is an urgent need. However, classifiers suffer from enormous dimensionality, since the dimensionality is determined by the number of distinct keywords in a document corpus. More seriously, most classifiers are either working slowly or they are constructed subjectively without any learning ability. In this paper, we address these problems with a fair feature-subset selection (FFSS) algorithm and an adaptive fuzzy learning network (AFLN) for classification. The FFSS algorithm is used to reduce the enormous dimensionality. It not only gives fair treatment to each category but also has ability to identify useful features, including both positive and negative features. On the other hand, the AFLN provides extremely fast learning ability to model the uncertain behavior for classification so as to correct the fuzzy matrix automatically. Experimental results show that both FFSS algorithm and the AFLN lead to a significant improvement in document classification, compared to alternative approaches.     

The correlation of texture and magnetic properties in Co-Cr films

This investigation focuses on the correlation between the crystallographic orientation of grains with respect to magnetic properties in Co-Cr films. Based on a distribution measured for the (0002) crystallographic pole densities, modified formulae describing the texture were used to calculate the coercivity Hc, the squareness ratio Rs (Mr⊥/TMs) and the anisotropy field Ha for Co-Cr films. In general, calculated values for the coercive force He agreed well with the measured values. The calculated Rs⊥. values were much larger than the measured ones, however, indicating that the influence of the demagnetizing field and the magnetostriction on the orientation of the magnetization cannot be neglected. On leave from Institute of Computing Technology, Academia Sinica, P.O.Box 2704-6, Beijing, P.R. China.     

Texture Changes During the Ripening of Port Salut Argentino Cheese in 2 Sampling Zones

ABSTRACT: Texture changes during ripening of Port Salut Argentino cheese for different sampling zones were studied. Compression relaxation tests were performed and results were analyzed using both Maxwellian and Peleg's models. Elastic equilibrium modulus obtained from the Maxwellian model decreased from 1.22 to 0.11 10⁴Pa during ripening. The constants derived from Peleg's model, k₁ and k₂, diminished with ripening time from 1.18 to 0.71 min and from 1.27 to 1.12, respectively. Asymptotic equilibrium modulus from Peleg's model decreased from 0.95 to 0.07 10⁴Pa during ripening. Rate parameters derived from a 1st order kinetics applied to both equilibrium moduli showed that the decrease was faster in the external zone (0.0846 d⁻¹) than in the central zone (0.0368 d⁻¹). The correlation between equilibrium moduli, salt concentration, moisture content, and maturation indexes was obtained with a determination coefficient of 0.76.     

Molecular correlations in sheared thermotropic copolyesters showing banded textures

Banded textures produced in a thermotropic liquid crystal polymer by shearing between glass slides are examined by using both transmission electron and polarized light microscopy. The periodic variation in director orientation about the shear axis, as measured by light microscopy, is shown to be distinctly different from that indicated by electron diffraction. Measurements of birefringence and observation of Zernicke phase contrast indicate periodic variations in optical properties of the polymer, in step with the bands. Such effects are accounted for in terms of a synchronous rotation of the planar aromatic groups about the molecular chain axes. Evidence for an out-of-plane component of molecular orientation is also presented.     

The microstructure of Cu films deposited by the self-ion assisted technique

The microstructures of Cu films deposited by the self-ion assisted, partially ionized beam (PIB) deposition technique under two different accelerating potentials, 0 KeV and 6 KeV, are compared. The 6 KeV film shows a bimodal (111) fiber and (100) fiber texture with an abundance of twin boundaries and a relatively large average grain size with a typical lognormal distribution. The 0 KeV film consists of small, mostly (111) oriented grains with islands of abnormally large (100) grains. The controlling factors for the abnormal growth of the (100) grains are discussed in relation to the observed microstructures, showing that all factors necessary for abnormal (100) growth are present in the films.