The increase in the number of monitor units in sliding window intensity-modulated radiotherapy, compared with conventional techniques for the same target dose, may lead to an increase in peripheral dose (PD). PD from a linear accelerator was measured for 6 MV X-ray using 0.6 cm3 ionization chamber inserted at 5 cm depth into a 35 cm x 35 cm x 105 cm plastic water phantom. Measurements were made for field sizes of 6 cm x 6 cm, 10 cm x 10 cm and 14 cm x 14 cm, shaped in both static and dynamic multileaf collimation (DMLC) mode, employing strip fields of fixed width 0.5 cm, 1.0 cm, 1.5 cm, and 2.0 cm, respectively. The effect of collimator rotation and depth of measurement on peripheral dose was investigated for 10 cm x 10 cm field. Dynamic fields require 2 to 14 times the number of monitor units than does a static open field for the same dose at the isocentre, depending on strip field width and field size. Peripheral dose resulting from dynamic fields manifests two distinct regions showing a crest and trough within 30 cm from the field edge and a steady exponential fall beyond 30 cm. All dynamic fields were found to deliver a higher PD compared with the corresponding static open fields, being highest for smallest strip field width and largest field size; also, the percentage increase observed was highest at the largest out-of-field distance. For 6 cm x 6 cm field, dynamic fields with 0.5 cm and 2 cm strip field width deliver PDs 8 and 2 times higher than that of the static open field. The corresponding factors for 14 cm x 14 cm field were 15 and 6, respectively. The factors by which PD for DMLC fields increase, relative to jaws-shaped static fields for out-of-field distance beyond 30 cm, are almost the same as the corresponding increases in the number of monitor units. Reductions of 20% and 40% in PD were observed when the measurements were done at a depth of 10 cm and 15 cm, respectively. When the multileaf collimator executes in-plane (collimator 90 degrees) motion, peripheral dose decreases by as much as a factor of 3 compared with cross-plane data. The knowledge of PD from DMLC field is necessary to estimate the increase in whole-body dose and the likelihood of radiation induced secondary malignancy.

COVID-19 is spreading widely across the globe right now. Majority of the countries are relying on models and studies such as stochastic simulations, AceMod model, neural networks-based models, exponential growth model,Weibull distribution model, and so on to forecast the number of COVID-19 cases in the coming months. The objective on utilizing these models is to ensure that strict measures can be enacted to contain the virus spread and also predict the resources required to deal with the pandemic as the disease spreads. In the past few months, several models were used to predict the infection rate for COVID-19. These models predicted the infection rates, recovery rate or death rates for the COVID-19 patients. All these different models took different approaches and different scenarios to predict the future rates. Now, that we know the real cases, we can check how accurate these models were. Some of these models were able to predict the near future quite close to the reality but most of them went astray. In this study, we review major forecasting models that were used in the context of India during country-wise lockdown and compare them. From these comparisons, we can see that while the advanced warning can be helpful in mitigating and preparing for an impending or ongoing epidemic, effects can be catastrophic if poorly fitting models are used for predictions.

No abstract

The complexity of interactions and the nature of the approximations made in the formulation of the algorithm require that the user be familiar with the limitations of various models. As computer power keeps growing, calculation algorithms are tending more towards physically based models. The nature and quantity of the data required varies according to the model which may be either measurement based models or physical based models. Multiple dose calculation algorithm support found in XiO Treatment Planning System can be used to advantage when choice is to be made between speed and accuracy. Thus XiO allows end users generate plans accurately and quickly to optimize the delivery of radiation therapy.

The traditional mode of recording faults in heavy factory equipment has been via handmarked inspection sheets, wherein a machine engineer manually marks the faulty machine regions on a paper outline of the machine. Over the years, millions of such inspection sheets have been recorded and the data within these sheets has remained inaccessible. However, with industries going digital and waking up to the potential value of fault data for machine health monitoring, there is an increased impetus towards digitization of these handmarked inspection records. To target this digitization, we propose a novel visual pipeline combining state of the art deep learning models, with domain knowledge and low level vision techniques, followed by inference of visual relationships. Our framework is robust to the presence of both static and nonstatic background in the document, variability in the machine template diagrams, unstructured shape of graphical objects to be identified and variability in the strokes of handwritten text. The proposed pipeline incorporates a capsule and spatial transformer network based classifier for accurate text reading, and a customized CTPN [15] network for text detection in addition to hybrid techniques for arrow detection and dialogue cloud removal. We have tested our approach on a real world dataset of 50 inspection sheets for large containers and boilers. The results are visually appealing and the pipeline achieved an accuracy of 87.1% for text detection and 94.6% for text reading.