A Deep-Fracture Approach to Embedded Multilayer Neural Networks for Handwritten Character Recognition
Downloads
Low-cost microcontrollers, but also with reduced processing and memory capabilities, are highly appreciated devices for the development of specific tasks as embedded systems. The devices found in the market today provide dedicated 32-bit architectures with communication capabilities ideal for IoT applications. Some of these tasks would highly benefit from the generalization and approximation capabilities of a neural network, but current convolutional networks prove to be too complex for microcontrollers. An alternative for some cases is the use of Multilayer Neural Networks (MNNs), which, thanks to a shallower depth, reach sizes suitable for use in small embedded systems. An MNN can accommodate a complex decision surface capable of being used in high dimensionality pattern classification problems, such as handwritten character recognition. Given the lack of an explicit strategy to define the architecture of these networks in coherence with the problem to be solved, this paper conducts a performance analysis of these networks in terms of their depth, to establish criteria to determine their size in a specific application. For the evaluation, the public database MNIST has been used in conjunction with classical metrics to evaluate the performance of a classification model. The results show that the depth of the network determines to a high degree the performance of the model, and guides the appropriate selection of the architecture.
Ao, X., Zhang, X.Y., & Liu, C.L., “Cross-modal prototype learning for zero-shot handwritten character recognition”, Pattern Recognition, 2022, 131(1):108859, ISSN 0031-3203, doi:10.1016/j.patcog.2022.108859.
Caley, J., Lawrance, N., & Hollinger, G., “Deep learning of structured environments for robot search”, Autonomous Robots, 2019, 43(7):1695–1714, ISSN 0929-5593,
doi:https://doi.org/10.1007/s10514-018-09821-4.
Castañeda, J. & Salguero, Y., “Adjustment of visual identification algorithm for use in stand-alone robot navigation applications”, Tekhnê, 2017, 14(1):73–86, ISSN 1692-8407.
Esquivel, J., Marín, N., & Martínez, F., “Digital development platform based on game boy advance and the ARM7 architecture”, Tekhnê, 2012, 9(1):5–12, ISSN 1692-8407.
Gajaria, D. & Adegbija, T., “Evaluating the performance and energy of STT-RAM caches for real-world wearable workloads”, Future Generation Computer Systems, 2022, 136(1):231–240, ISSN 0167-739X, doi:10.1016/j.future.2022.05.023.
Hashemifar, Z., Adhivarahan, C., Balakrishnan, A., & Dantu, K., “Augmenting visual slam with wi-fi sensing for indoor applications”, Autonomous Robots, 2019, 43(8):2245–2260,ISSN 0929-5593, doi:https://doi.org/10.1007/s10514-019-09874-z.
Hock, A. & Schoelling, A., “Distributed iterative learning control for multi-agent systems”, Autonomous Robots, 2019, 43(8):1989–2010, ISSN 0929-5593, doi:https://doi.org/10.1007/s10514-019-09845-4.
Hu, J., Wang, P., Xu, C., Zhou, H., & Yao, J., “High accuracy adaptive motion control for a robotic manipulator with model uncertainties based on multilayer neural network”, Asian Journal of Control, 2021, 24(3):1503–1514, ISSN 1561-8625, doi:10.1002/asjc.2546.
Ktari, J., Frikha, T., Hamdi, M., Elmannai, H., & Hmam, H., “Lightweight AI framework for industry 4.0 case study: Water meter recognition”, Big Data and Cognitive Computing, 2022, 6(3):72, ISSN 2504-2289, doi:10.3390/bdcc6030072.
Martínez, F., Jacinto, E., & Montiel, H., “Neuronal environmental pattern recognizer: Optical-by-distance LSTM model for recognition of navigation patterns in unknown environments”, Communications in Computer and Information Science, 2019, 1071(1):220–227, ISSN 1865-0929,doi:https://doi.org/10.1007/978-981-32-9563-6_23.
Martínez, F., Penagos, C., & Pacheco, L., “Deep regression models for local interaction in multi-agent robot tasks”, Lecture Notes in Computer Science, 2018a, 10942(1):66–73, ISSN 0302-9743, doi:10.1007/978-3-319-93818-9_7.
Martínez, F. & Rendón, A., “Implementation and evaluation of advantage actor-critic algorithm on a desktop computer with a multi-core CPU”, International Journal of Computer Applications Technology and Research, 2022, 11(07):284–290, ISSN 2319-8656, doi:10.7753/ijcatr1107.1005.
Martínez, F., Rendón, A., & Arbulú, M., “A data-driven path planner for small autonomous robots using deep regression models”, Lecture Notes in Computer Science, 2018b, 10943(1):596–603, ISSN 0302-9743, doi:https://doi.org/10.1007/978-3-319-93803-5_56.
Moreno, A. & Páez, D., “Performance evaluation of ros on the raspberry pi platform as os for small robots”, Tekhnê, 2017, 14(1):61–72, ISSN 1692-8407.
Prabakaran, G., Vaithiyanathan, D., & Ganesan, M., “FPGA based intelligent embedded system for predicting the productivity using fuzzy logic”, Sustainable Computing: Informatics and Systems, 2022, 35(1):100749, ISSN 2210-5379,
doi:10.1016/j.suscom.2022.100749.
Riaño, J., Ladino, C., & Martínez, F., “Implementation of fft transform on a FPGA oriented towards its application in power electronic converters”, Tekhnê, 2012, 9(1):21–32, ISSN 1692-8407.
Shi, W., Huang, Z., Huang, H., Hu, C., Chen, M., Yang, S., & Chen, H., “LOEN: Lensless opto-electronic neural network empowered machine vision”, Light: Science & Applications, 2022, 11(1):121, ISSN 2095-5545, doi:10.1038/s41377-022-00809-5.
Xu, F., Guo, Z., Chen, H., Ji, D., & Qu, T., “A custom parallel hardware architecture of nonlinear model-predictive control on FPGA”, IEEE Transactions on Industrial Electronics, 2022, 69(11):11569–11579, ISSN 0278-0046, doi:10.1109/tie.2021.3118427.
Yao, Y., Liu, S., Wu, S., Ni, J., Wang, J., Yang, G., & Zhang, Y., “WAMP22S: workload-aware GPU performance model based pseudo-preemptive real-time scheduling for the airborne embedded system”, IEEE Transactions on Parallel and Distributed Systems, 2021, 33(11):2767–2780, ISSN 1045-9219,
doi:10.1109/tpds.2021.3134269.
Yarza, I., Agirre, I., Mugarza, I., & Cerrolaza, J.P., “Safety and security collaborative analysis framework for high-performance embedded computing devices”, Microprocessors and Microsystems, 2022, 93(1):104572, ISSN 0141-9331,
doi:10.1016/j.micpro.2022.104572.
Zilong, H., Jinshan, T., Ziming, W., Kai, Z., Ling, Z., & Qingling, S., “Deep learning for image-based cancer detection and diagnosis - a survey”, Pattern Recognition, 2018, 83(1):134–149, ISSN 0031-3203, doi:https://doi.org/10.1016/j.patcog.2018.05.014.
