Comparative analysis of image projection-based descriptors in Siamese neural networks. (April 2021)

• Record Type: Journal Article
• Authors: Kertész, Gábor; Szénási, Sándor; Vámossy, Zoltán
• Is Part Of: Advances in engineering software. Volume 154(2021)
• Publication Date: 2021-04
• Subjects: Siamese neural networks -- Multi-directional image projections -- Neural architecture generation -- one-shot classification accuracy -- Vehicle re-identification; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2020.102963
• Languages: English
• ISSNs: 0965-9978

Abstract

Highlights: Siamese neural networks were trained on image projection matrices for a vehicle re-identification task. A neural architecture generation method based on backtracking search is used for convolutional architecture generation. The distributed training is done in a master/worker structure with LPT-heuristics based scheduling on estimated complexity. Projection-based methods are Pareto optimal in terms of one-shot classification accuracy and memory consumption. Abstract: Low-level object matching can be done using projection signatures. In case of a large number of projections, the matching algorithm has to deal with less significant slices. A trivial approach would be to do statistical analysis or apply machine learning to determine the significant features. To take adjacent values of the projection matrices into account, a convolutional neural network should be used. To compare two matrices, a Siamese-structure of convolutional heads can be applied. In this paper, an experiment is designed and implemented to analyze the object matching performance of Siamese Convolutional Neural Networks based on multi-directional image projection data. A backtracking search-based Neural Architecture Generation method is used to create convolutional architectures, and a Master/Worker structured distributed processing with highly efficient scheduling based on the Longest Processing Times-heuristics is used for parallel training and evaluation of the models. Results show that the projection-based methods are Pareto optimal in terms of one-shot classification accuracy and memory consumption.