This is an old revision of the document!
Table of Contents
Рабочие документы
См. также раздел "Ограниченный доступ"
Общие руководства по МО
- Understanding Deep Learning, Simon J.D. Prince
Хорошее введение в МО с математическими подробнотями.
Форматы данных
Анализ изображений IACT
Параметры Хилласа
Jashwanth S2 , Sudeep Ghosh1 , Kavitha Yogaraj1 , Neha Shah2 , and Ankhi Roy3. “G AMMA - HADRON S EPARATION IN I MAGING ATMOSPHERIC C HERENKOV T ELESCOPES USING Q UANTUM C LASSIFIERS”. ArXiv:2210.03771
2.2 Image parametrization
The hadronic shower images are observed to be more longer and broader compared to gamma ray shower images. Effective discrimination of primary gamma ray shower and hadronic background shower is possible on the basis of the width,length and orientation of these images. The showers with axis parallel to the optic axis and landing directly on the detector will produce circular images. But if it lands at some distance(impact parameter) away from the detector it will be a bivariate Gaussian distribution which is an elliptical cluster. For gamma ray showers the major axis is oriented towards the camera center whereas the hadronic showers have random axial orientation. The pixel image of the shower after some pre processing and image cleaning is then converted to into few image parameters, defined by moment analysis on the pixel signal amplitudes. The moments are defined as:
(1)
where, x and y are the coordinates of pixels and n is the number of digital counts in a pixel. The summation runs over all the pixels in the image.
Figure 1: Definitions of some hillas parameters in the camera plane.
Image spreads are derived from the moments in (1):
(2)
Image parameters can be derived from the moments and the image spreads in (2):
(3)
where, miss is the perpendicular distance between camera center and the major axis as shown in Fig. 1.
Multimodal
Нормализующие потоки
- См. также Understanding Deep Learning
Domain Adaptation
Стерео режим и мультимодальные данные
Восстановление энергии
Прочее
ПО и другие вопросы
- Руководства по различным ПО
- Установка Jupyter на Ubuntu
- Установка TensorFlow на Ubuntu
Благодарности
Исследование выполнено в рамках государственного задания Московского государственного университета имени М.В. Ломоносова с использованием оборудования, предоставленного по Программе развития МГУ, и данных, полученных на уникальной эксперментальной установке ТАЙГА. Исследование выполнено за счет средств гранта Российского научного фонда № 24-11-00136, https://rscf.ru/project/24-11-00136/ ------------- The study was conducted under the state assignment of Lomonosov Moscow State University using equipment provided under the MSU Development Program and data obtained on the unique experimental setup TAIGA. The study was carried out at the expense of the grant of the Russian Science Foundation No. 24-11-00136, https://rscf.ru/project/24-11-00136/
