Explore chapters and articles related to this topic
Data Serialization and Deserialization
Published in Hossam Fattah, LTE™ Cellular Narrowband Internet of Things (NB-IoT), 2021
Json serialization (or deserialization) is the process of translating a data structure or object into JSON formatted text or convert JSON formatted text back to a data structure or object. The following data structure when serialized, results in the JSON-formatted text. When deserialized, the JSON-formatted text results in the data structure.
An improved memory networks based product model classification method
Published in International Journal of Computer Integrated Manufacturing, 2021
Chengfeng Jian, Lingming Liang, Keyi Qiu, Meiyu Zhang
Aiming at the above problems, a classification and recognition method based on improved LSTM model (ILSTM) is proposed. First, the B-rep information is extracted from the MBD model and preprocess it into the edge-side attribute adjacency graph. Then, this paper used a model attribute adjacency graph serialization method based on adjacency matrix to map the geometry, topology and attribute information of the model to sequence coding of unequal length. Finally, the processed model sequence coding is classified and identified by the improved LSTM. In order to improve the discriminative power of the model, a new loss function (M-Softmax loss) is designed during the training process. At the same time, adaptive training is proposed to ensure the convergence of the model.
Optimizing segmented trajectory data storage with HBase for improved spatio-temporal query efficiency
Published in International Journal of Digital Earth, 2023
Yi Bao, Zhou Huang, Xuri Gong, Yuyang Zhang, Ganmin Yin, Han Wang
The trajectory storage and management, as the basis of computation and mining, can provide the function of trajectory data storage and basic query, and plays an important role in various fields of trajectory data application. We have investigated the trajectory segments storage method and its associated optimization techniques in a distributed NoSQL database, covering aspects such as segmentation, indexing, serialization, query optimization, and calculation. Based on the HBase database, we implement the prototype system of segmented trajectory storage, verify the effect of various designs and optimizations with real data, and compare it with Geomesa to show the advantages and characteristics of the segmented storage model.
Scalable Feature Tracking for Finite Element Meshes Demonstrated with a Novel Phase-Field Grain Subdivision Model
Published in Nuclear Technology, 2021
Cody J. Permann, Andrea M. Jokisaari, Michael R. Tonks, Daniel Schwen, Derek R. Gaston, Fande Kong, Robert Hiromoto, Richard C. Martineau
After all partial features are post processed on all processor ranks, the data structure is ready for serialization. The subset of data mentioned in Sec. II.B is serialized into raw variable length byte arrays and packed into num_variables variable-length binary buffers, one for each variable (order parameter) used in the simulation per processor. The binary buffers themselves contain embedded length indicators for use during the deserialization process. These feature records do not contain any of the local (interior) IDs, which drastically reduces the size of these serialized structures.