scholarly journals Classification of Ice Crystal Habits Observed From Airborne Cloud Particle Imager by Deep Transfer Learning

2019 ◽  
Vol 6 (10) ◽  
pp. 1877-1886 ◽  
Author(s):  
Haixia Xiao ◽  
Feng Zhang ◽  
Qianshan He ◽  
Pu Liu ◽  
Fei Yan ◽  
...  
Keyword(s):  
Transfer Learning ◽  
Ice Crystal ◽  
Cloud Particle ◽  
Ice Crystal Habits

A Classification of Ice Crystal Habits Using Combined Lidar and Scanning Polarimeter Observations during the SEAC4RS Campaign

2020 ◽  
Vol 37 (12) ◽  
pp. 2185-2196
Author(s):  
Natalie Midzak ◽  
John E. Yorks ◽  
Jianglong Zhang ◽  
Bastiaan van Diedenhoven ◽  
Sarah Woods ◽  
...  
Keyword(s):  
Atmospheric Composition ◽  
Crystal Habit ◽  
Ice Crystal ◽  
Cloud Physics ◽  
Cloud Particle ◽  
Clustering Technique ◽  
Ice Crystal Habits

AbstractUsing collocated NASA Cloud Physics Lidar (CPL) and Research Scanning Polarimeter (RSP) data from the Studies of Emissions and Atmospheric Composition, Clouds and Climate Coupling by Regional Surveys (SEAC4RS) campaign, a new observational-based method was developed which uses a K-means clustering technique to classify ice crystal habit types into seven categories: column, plates, rosettes, spheroids, and three different type of irregulars. Intercompared with the collocated SPEC, Inc., Cloud Particle Imager (CPI) data, the frequency of the detected ice crystal habits from the proposed method presented in the study agrees within 5% with the CPI-reported values for columns, irregulars, rosettes, and spheroids, with more disagreement for plates. This study suggests that a detailed ice crystal habit retrieval could be applied to combined space-based lidar and polarimeter observations such as CALIPSO and POLDER in addition to future missions such as the Aerosols, Clouds, Convection, and Precipitation (A-CCP).

A Comprehensive Habit Diagram for Atmospheric Ice Crystals: Confirmation from the Laboratory, AIRS II, and Other Field Studies

2009 ◽  
Vol 66 (9) ◽  
pp. 2888-2899 ◽  
Author(s):  
Matthew P. Bailey ◽  
John Hallett
Keyword(s):  
Field Studies ◽  
Crystal Habit ◽  
Ice Crystal ◽  
Comprehensive Description ◽  
Cloud Particle ◽  
In Situ Observations ◽  
Ice Crystal Habits ◽  
Habit Modification ◽  
Recent Laboratory

Abstract Recent laboratory experiments and in situ observations have produced results in broad agreement with respect to ice crystal habits in the atmosphere. These studies reveal that the ice crystal habit at −20°C is platelike, extending to −40°C, and not columnar as indicated in many habit diagrams found in atmospheric science journals and texts. These diagrams were typically derived decades ago from laboratory studies, some with inherent habit bias, or from combinations of laboratory and in situ observations at the ground, observations that often did not account for habit modification by precipitation from overlying clouds of varying temperatures. Habit predictions from these diagrams often disagreed with in situ observations at temperatures below −20°C. More recent laboratory and in situ studies have achieved a consensus on atmospheric ice crystal habits that differs from the traditional habit diagrams. These newer results can now be combined to give a comprehensive description of ice crystal habits for the atmosphere as a function of temperature and ice supersaturation for temperatures from 0° to −70°C, a description dominated by irregular and imperfect crystals. Cloud particle imager (CPI) habit observations made during the Second Alliance Icing Research Study (AIRS II) and elsewhere corroborate this comprehensive habit description, and a new habit diagram is derived from these results.

scholarly journals Ice Crystal Linear Growth Rates from −20° to −70°C: Confirmation from Wave Cloud Studies

2012 ◽  
Vol 69 (1) ◽  
pp. 390-402 ◽  
Author(s):  
Matthew Bailey ◽  
John Hallett
Keyword(s):  
Growth Rates ◽  
Volume Growth ◽  
Field Studies ◽  
Atmospheric Conditions ◽  
Ice Crystal ◽  
Cloud Particle ◽  
Simple Method ◽  
Laboratory Results ◽  
Ice Crystal Habits

Abstract As a result of recent comprehensive laboratory and field studies, many details have been clarified concerning atmospheric ice crystal habits below −20°C as a function of temperature, ice supersaturation, air pressure, and growth history. A predominance of complex shapes has been revealed that is not reflected in most models, with symmetric shapes often incorrectly emphasized. From the laboratory study, linear (maximum dimension), projected area, and volume growth rates of complex and simple habits have been measured under simulated atmospheric conditions for temperatures from −20° to −70°C. Presently, only a few in situ cases of measured ice crystal growth rates are available for comparison with laboratory results. Observations from the Interaction of Aerosol and Cold Clouds (INTACC) field study of a well-characterized wave cloud at −24°C are compared with the laboratory results using a simple method of habit averaging to derive a range of expected growth rates. Laboratory results are also compared with recently reported wave cloud results from the Ice in Clouds Experiment–Layer Clouds (ICE-L) study between −20° and −32°C, in addition to a much colder wave cloud at −65°C. Considerable agreement is observed in these cases, confirming the reliability of the laboratory measurements. This is the first of two companion papers that compare laboratory growth rates and characteristics with in situ measurements, confirming that the laboratory results effectively provide a predictive capability for cloud particle and particle ensemble growth.

scholarly journals Transfer learning networks with skip connections for classification of brain tumors

2021 ◽  
Author(s):  
Saleh Alaraimi ◽  
Kenneth E. Okedu ◽  
Hugo Tianfield ◽  
Richard Holden ◽  
Omair Uthmani
Keyword(s):  
Brain Tumors ◽  
Transfer Learning ◽  
Learning Networks

scholarly journals Classification of jujube defects in small data sets based on transfer learning

2021 ◽  
Author(s):  
Jianping Ju ◽  
Hong Zheng ◽  
Xiaohang Xu ◽  
Zhongyuan Guo ◽  
Zhaohui Zheng ◽  
...  
Keyword(s):  
Transfer Learning ◽  
Loss Function ◽  
Training Model ◽  
Parameter Distribution ◽  
Test Accuracy ◽  
Small Data ◽  
Data Sets ◽  
Data Set ◽  
Small Data Sets

AbstractAlthough convolutional neural networks have achieved success in the field of image classification, there are still challenges in the field of agricultural product quality sorting such as machine vision-based jujube defects detection. The performance of jujube defect detection mainly depends on the feature extraction and the classifier used. Due to the diversity of the jujube materials and the variability of the testing environment, the traditional method of manually extracting the features often fails to meet the requirements of practical application. In this paper, a jujube sorting model in small data sets based on convolutional neural network and transfer learning is proposed to meet the actual demand of jujube defects detection. Firstly, the original images collected from the actual jujube sorting production line were pre-processed, and the data were augmented to establish a data set of five categories of jujube defects. The original CNN model is then improved by embedding the SE module and using the triplet loss function and the center loss function to replace the softmax loss function. Finally, the depth pre-training model on the ImageNet image data set was used to conduct training on the jujube defects data set, so that the parameters of the pre-training model could fit the parameter distribution of the jujube defects image, and the parameter distribution was transferred to the jujube defects data set to complete the transfer of the model and realize the detection and classification of the jujube defects. The classification results are visualized by heatmap through the analysis of classification accuracy and confusion matrix compared with the comparison models. The experimental results show that the SE-ResNet50-CL model optimizes the fine-grained classification problem of jujube defect recognition, and the test accuracy reaches 94.15%. The model has good stability and high recognition accuracy in complex environments.

A novel transfer learning approach for the classification of histological images of colorectal cancer

2021 ◽  
Author(s):  
Elene Firmeza Ohata ◽  
João Victor Souza das Chagas ◽  
Gabriel Maia Bezerra ◽  
Mohammad Mehedi Hassan ◽  
Victor Hugo Costa de Albuquerque ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Transfer Learning ◽  
Learning Approach ◽  
Histological Images

scholarly journals Transfer Learning with CNN for Classification of Weld Defect

IEEE Access ◽  
2021 ◽  
pp. 1-1
Author(s):  
Samuel Kumaresan ◽  
K. S. Jai Aultrin ◽  
S. S. Kumar ◽  
M. Dev Anand
Keyword(s):  
Transfer Learning ◽  
Weld Defect
Sign in / Sign up

Export Citation Format

Share Document