基于F3Net显著性目标检测的蝴蝶图像前背景自动分割

[目的]具有复杂背景的蝴蝶图像前背景分割难度大.本研究旨在探索基于深度学习显著性目标检测的蝴蝶图像自动分割方法.[方法]应用DUTS-TR数据集训练F3Net显著性目标检测算法构建前背景预测模型,然后将模型用于具有复杂背景的蝴蝶图像数据集实现蝴蝶前背景自动分割.在此基础上,采用迁移学习方法,保持ResNet骨架不变,利...     

How biologists conceptualize genes: an empirical study

Philosophers and historians of biology have argued that genes are conceptualized differently in different fields of biology and that these differences influence both the conduct of research and the interpretation of research by audiences outside the field in which the research was conducted. In this paper we report the results of a questionnaire study of how genes are conceptualized by biological scientists at the University of Sydney, Australia. The results provide tentative support for some hypotheses about conceptual differences between different fields of biological research.     

Mouse knockout of guanylyl cyclase C: Recognition memory deficits in the absence of activity changes

Guanylyl cyclase C (GC‐C) is found in brain regions where dopamine is expressed. We characterized a mouse in which GC‐C was knocked out (KO) that was reported to be a model of attention deficit hyperactivity disorder (ADHD). We re‐examined this model and controlled for litter effects, used 16 to 23 mice per genotype per sex and assessed an array of behavioral and neurochemical outcomes. GC‐C KO mice showed no phenotypic differences from wild‐type mice on most behavioral tests, or on striatal or hippocampal monoamines, and notably no evidence of an ADHD‐like phenotype. KO mice were impaired on novel object recognition, had decreased tactile startle but not acoustic startle, and females had increased latency on cued training trials in the Morris water maze, but not hidden platform spatial learning trials. Open‐field activity showed small differences in females but not males. The data indicate that the GC‐C KO mouse with proper controls and sample sizes has a moderate cognitive and startle phenotype but has no ADHD‐like phenotype.     

Object classification by echolocation in nectar feeding bats: size-independent generalization of shape

The nectar-feeding bat Glossophaga can be trained to discriminate two hollow forms, a hollow hemisphere and a paraboloid with the same diameter and depth, in total darkness. During training a saturation level of about 85-90% correct choices or more can be reached within 50-100 visits. To investigate generalization abilities, the bats were tested with pairs of the same shape but of different size. Although no reward was offered, they preferred the hollow sphere (30 mm and 50 mm diameter, but not 18 mm) over the corresponding paraboloids. Thus, the bats were able to generalize some features of the rewarded form and detect them in forms of the same shape but different size. This transposition is remarkable, since the bats could not use absolute spectral characters, but had to pay attention to size-independent features common to hollow hemispheres. Possible cues are the variation of echoes in dependence of different angles of calling direction (constant in spheres, changing with position in paraboloids) and/or the "timbre" of the echoes, i.e. their spectral pattern independent of their absolute pitch     

The physics of symbols: bridging the epistemic cut

Evolution requires the genotype–phenotype distinction, a primeval epistemic cut that separates energy-degenerate, rate-independent genetic symbols from the rate-dependent dynamics of construction that they control. This symbol–matter or subject–object distinction occurs at all higher levels where symbols are related to a referent by an arbitrary code. The converse of control is measurement in which a rate-dependent dynamical state is coded into quiescent symbols. Non-integrable constraints are one necessary condition for bridging the epistemic cut by measurement, control, and coding. Additional properties of heteropolymer constraints are necessary for biological evolution.     

Multifrequency Microwave Imaging for Brain Stroke Detection

CT and MRI are often used in the diagnosis and monitoring of stroke. However, they are expensive, time-consuming, produce ionizing radiation (CT), and not suitable for continuous monitoring stroke. Microwave imaging (MI) has been extensively investigated for identifying several types of human organs, including breast, brain, lung, liver, and gastric. The authors recently developed a holographic microwave imaging (HMI) algorithm for biological object detection. However, this method has difficulty in providing accurate information on embedded small inclusions. This paper describes the feasibility of the use of a multifrequency HMI algorithm for brain stroke detection. A numerical system, including HMI data collection model and a realistic head model, was developed to demonstrate the proposed method for imaging of brain tissues. Various experiments were carried out to evaluate the performance of the proposed method. Results of experiments carried out using multifrequency HMI have been compared with the results obtained from single frequency HMI. Results showed that multifrequency HMI could detect strokes and provide more accurate results of size and location than the single frequency HMI algorithm.     

Cracking the nut: geographical adjacency of sister taxa supports vicariance in a polytomic salamander clade in the absence of node support

The urodelan genus Lyciasalamandra, which inhabits a relatively small area along the southern Turkish coast and some Aegean islands, provides an outstanding example of a diverse but phylogenetically unresolved taxon. Molecular trees contain a single basal polytomy that could be either soft or hard. We here use the information of nuclear (allozymes) and mitochondrial (fractions of the 16S rRNA and ATPase genes) datasets in combination with area relationships of lineages to resolve the phylogenetic relationships among Lyciasalamandra species in the absence of sufficient node support. We can show that neither random processes nor introgressive hybridization can be invoked to explain that the majority of pairs of sister taxa form geographically adjacent units and interpret that this pattern has been shaped by vicariant events. Topology discordance between mitochondrial and nuclear trees mainly refers to an affiliation of L. helverseni, a taxon restricted to the Karpathos archipelago, to the western-most and geographically proximate mainland taxon in the nuclear tree, while in the organelle tree it turns out to be the sister lineage to the geographically most distant eastern clade. As this discordance cannot be explained by long-branch attraction in either dataset we suppose that oversea dispersal may have accounted for a second colonization of the Karpathos archipelago. It may have initiated introgression and selection driven manifestation of alien eastern mitochondrial genomes on a western nuclear background. Our approach of testing for area relationships of sister taxa against the null hypothesis of random distribution of these taxa seems to be especially helpful in phylogenetic studies where traditional measures of phylogenetic branch support fail to reject the null hypothesis of a hard polytomy.