China
Africa
Explore chapters and articles related to this topic
Data Formats
Published in Julio Sanchez, Maria P. Canton, William Perrizo, Space Image Processing, 2018
Julio Sanchez, Maria P. Canton, William Perrizo
The total byte size of the FITS file is the number of header units, times 2880 bytes, plus the byte size of the data array. This actually refers to the meaningful data since the data array is padded with ASCII blanks in order to make it a multiple of 2880 bytes. In the case of Figure 8.2, if there is a single header unit, the size of the FITS data file is
DNN based approach to classify Covid’19 using convolutional neural network and transfer learning
Published in International Journal of Computers and Applications, 2022
Bhavya Joshi, Akhilesh Kumar Sharma, Narendra Singh Yadav, Shamik Tiwari
Precision or Positive prediction value (PPV) shows the amount of positive samples which are classified correctly divided by the sum of false positive and true positive samples [41]. Recall, Sensitivity, True positive rate (TPR) or hit rate of a classifier shows the positive correctly classified samples divided by the sum of false negative and true positive samples [41]. If there are too many False Positives, then the recall will be low. Inverse Recall, Specificity, or True negative rate (TNR) is expressed as the proportion of the negative samples that are classified correctly divided by all the negative samples. Thus, the specificity shows the proportion of the negative samples which got correctly classified [41,47]. The performance of the proposed CNN as well as both the Deep Transfer networks is evaluated using the aforementioned metrics on the same dataset. All three models are trained for multiple epochs and the training curves are plotted to evaluate the training. Figure 6 shows the two curves for the CNN. The model fits well into the dataset and is able to find the patterns in the X-Ray images. The model's confusion matrix is also important and thus plotted.
Thermal decomposition of struvite in water: qualitative and quantitative mineralogy analysis
Published in Environmental Technology, 2020
Athanasius P. Bayuseno, Wolfgang W. Schmahl
Before undertaking an analysis of the heated samples, X-ray diffraction patterns of the pure struvite (S), and the empty sample holder were examined (Figure 1). The X-ray diffraction pattern of the S powder sample shows the characteristic peaks corresponding to those of the struvite standard pattern in the powder diffraction file [PDF#71-2089]. The results of the XRD Rietveld analysis also confirmed that the X-ray profile pattern of the crystalline solid fits very well with that of the struvite model as shown in Figure 1(b). Struvite crystallizes in the orthorhombic system (space group Pm21n) with the lattice parameters a = 6•9335 Å, b = 6•1422 Å, and c = 11•1784 Å [27]. Moreover, the Kapton® Polyimide films of the empty sample holder shows no significant signals of amorphous phase [28]. After collecting the X-ray diffraction patterns of the unheated samples, the heated samples were directly measured by XRD and the diffraction patterns are presented in Figure 2.
Network topology of deeply supercooled water
Published in Molecular Physics, 2019
C. Shi, O. L. G. Alderman, C. J. Benmore
EPSR fits to the high-energy x-ray diffraction structure factors of water at 293, 262, 255 and 244 K are shown in Figure 1(a). Neutron diffraction data on supercooled colloidal D2O published by Bellissent-Funel et al. [21], and measured at similar temperatures to our x-ray diffraction data, are compared to predictions from our models in Figure 1(b). These neutron data were not used as constraints in our EPSR fits, but are shown as a consistency check to verify that the water models are reasonable. Selected O-O differential pair distribution functions (multiplied by r to emphasise correlations at the longer distances) are shown in Figure 2. At room temperature our water model is shown to be in good agreement with that published by Soper [46]. The slight difference observed between this study and [46] is most likely because of the input data. The existing fit was heavily weighted towards neutron data and the hydrogen correlations, whereas this study was based entirely on high-quality x-ray data (dominated by the O-O interactions) measured out to high-Q values. All peaks in r.dOO(r) increase in intensity and sharpen as water supercools, and at the lowest temperature of 244 K we observe distinct peaks in the O-O network at 2.8, 4.5, 6.8, 8.8 and 10.8 Å.