# A Hand Gesture Recognition Circuit Utilizing an Analog Voting Classifier

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Background

#### 2.1. Electromyography and Applications

#### 2.2. Mathematical Modeling of Centroid-Based Classifications

## 3. Proposed Architecture

#### 3.1. Centroid-Based Classifier

#### 3.2. Voting Circuit

## 4. Application Example and Simulation Results

## 5. Performance Summary and Discussion

## 6. Conclusions

## Author Contributions

## Funding

## Data Availability Statement

## Conflicts of Interest

## References

- Iqbal, M.H.; Aydin, A.; Brunckhorst, O.; Dasgupta, P.; Ahmed, K. A review of wearable technology in medicine. J. R. Soc. Med.
**2016**, 109, 372–380. [Google Scholar] [CrossRef] - Kim, J.; Campbell, A.S.; de Ávila, B.E.F.; Wang, J. Wearable biosensors for healthcare monitoring. Nat. Biotechnol.
**2019**, 37, 389–406. [Google Scholar] [CrossRef] [PubMed] - Sharma, A.; Badea, M.; Tiwari, S.; Marty, J.L. Wearable biosensors: An alternative and practical approach in healthcare and disease monitoring. Molecules
**2021**, 26, 748. [Google Scholar] [CrossRef] - Bonato, P. Advances in wearable technology and its medical applications. In Proceedings of the IEEE 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology, Buenos Aires, Argentina, 31 August–4 September 2010; pp. 2021–2024. [Google Scholar]
- Chan, M.; Estève, D.; Fourniols, J.Y.; Escriba, C.; Campo, E. Smart wearable systems: Current status and future challenges. Artif. Intell. Med.
**2012**, 56, 137–156. [Google Scholar] [CrossRef] [PubMed] - Li, Y.; Poon, C.C.; Zhang, Y.T. Analog integrated circuits design for processing physiological signals. IEEE Rev. Biomed. Eng.
**2010**, 3, 93–105. [Google Scholar] [CrossRef] [PubMed] - Shi, W.; Dustdar, S. The promise of edge computing. Computer
**2016**, 49, 78–81. [Google Scholar] [CrossRef] - Haensch, W.; Gokmen, T.; Puri, R. The next generation of deep learning hardware: Analog computing. Proc. IEEE
**2018**, 107, 108–122. [Google Scholar] [CrossRef] - Özdemir, A.T.; Barshan, B. Detecting falls with wearable sensors using machine learning techniques. Sensors
**2014**, 14, 10691–10708. [Google Scholar] [CrossRef] [PubMed] - Ehrmann, G.; Blachowicz, T.; Homburg, S.V.; Ehrmann, A. Measuring biosignals with single circuit boards. Bioengineering
**2022**, 9, 84. [Google Scholar] [CrossRef] [PubMed] - Farina, D.; Merletti, R.; Enoka, R.M. The extraction of neural strategies from the surface EMG. J. Appl. Physiol.
**2004**, 96, 1486–1495. [Google Scholar] [CrossRef] [PubMed] - Lo Presti, D.; Carnevale, A.; D’Abbraccio, J.; Massari, L.; Massaroni, C.; Sabbadini, R.; Schena, E. A multi-parametric wearable system to monitor neck movements and respiratory frequency of computer workers. Sensors
**2020**, 20, 536. [Google Scholar] [CrossRef] [PubMed][Green Version] - Hong, Y.; Wang, B.; Lin, W.; Jin, L.; Liu, S.; Luo, X.; Yang, Z. Highly anisotropic and flexible piezoceramic kirigami for preventing joint disorders. Sci. Adv.
**2021**, 7, eabf0795. [Google Scholar] [CrossRef] [PubMed] - Gu, Y.; Zhang, T.; Chen, H.; Wang, F.; Pu, Y.; Gao, C.; Li, S. Mini review on flexible and wearable electronics for monitoring human health information. Nanoscale Res. Lett.
**2019**, 14, 1–15. [Google Scholar] [CrossRef] [PubMed][Green Version] - Nasiri, S.; Khosravani, M.R. Progress and challenges in fabrication of wearable sensors for health monitoring. Sens. Actuators A Phys.
**2020**, 312, 112105. [Google Scholar] [CrossRef] - Kakria, P.; Tripathi, N.K.; Kitipawang, P. A real-time health monitoring system for remote cardiac patients using smartphone and wearable sensors. Int. J. Telemed. Appl.
**2015**, 2015, 8. [Google Scholar] [CrossRef][Green Version] - Wang, A.; Calhoun, B.H.; Chandrakasan, A.P. Sub-Threshold Design for Ultra Low-Power Systems (Volume 95); Springer: New York, NY, USA, 2006. [Google Scholar]
- Reid, R.C.; Mahbub, I. Wearable self-powered biosensors. Curr. Opin. Electrochem.
**2020**, 19, 55–62. [Google Scholar] [CrossRef] - Grattieri, M.; Minteer, S.D. Self-powered biosensors. ACS Sens.
**2018**, 3, 44–53. [Google Scholar] [CrossRef][Green Version] - Alimisis, V.; Gourdouparis, M.; Gennis, G.; Dimas, C.; Sotiriadis, P.P. Analog Gaussian Function Circuit: Architectures, Operating Principles and Applications. Electronics
**2021**, 10, 2530. [Google Scholar] [CrossRef] - Lazzaro, J.; Ryckebusch, S.; Mahowald, M.A.; Mead, C.A. Winner-take-all networks of O (n) complexity. Adv. Neural Inf. Process. Syst.
**1988**, 1, 703. [Google Scholar] - Starter: Electromyography(EMG) Dataset 533d170a-8. Available online: https://www.kaggle.com/code/kerneler/starter-electromyography-emg-dataset-533d170a-8/notebook (accessed on 29 October 2022).
- Hung, H.; Adzic, V. Monte carlo simulation of device variations and mismatch in analog integrated circuits. Proc. NCUR
**2006**, 2006, 1–8. [Google Scholar] - Martinek, R.; Ladrova, M.; Sidikova, M.; Jaros, R.; Behbehani, K.; Kahankova, R.; Kawala-Sterniuk, A. Advanced Bioelectrical Signal Processing Methods: Past, Present, and Future Approach—Part III: Other Biosignals. Sensors
**2021**, 21, 6064. [Google Scholar] [CrossRef] - Reaz, M.B.I.; Hussain, M.S.; Mohd-Yasin, F. Techniques of EMG signal analysis: Detection, processing, classification and applications. Biol. Proced. Online
**2006**, 8, 11–35. [Google Scholar] [CrossRef] [PubMed] - Hogrel, J.Y. Clinical applications of surface electromyography in neuromuscular disorders. Neurophysiol. Clin. Neurophysiol.
**2005**, 35, 59–71. [Google Scholar] [CrossRef] [PubMed] - Sudarsan, S.; Sekaran, E.C. Design and development of EMG controlled prosthetics limb. Procedia Eng.
**2012**, 38, 3547–3551. [Google Scholar] [CrossRef][Green Version] - Castellini, C.; Van Der Smagt, P. Surface EMG in advanced hand prosthetics. Biol. Cybern.
**2009**, 100, 35–47. [Google Scholar] [CrossRef] [PubMed][Green Version] - Yousefi, J.; Hamilton-Wright, A. Characterizing EMG data using machine-learning tools. Comput. Biol. Med.
**2014**, 51, 1–13. [Google Scholar] [CrossRef] [PubMed] - Saridis, G.N.; Gootee, T.P. EMG pattern analysis and classification for a prosthetic arm. IEEE Trans. Biomed. Eng.
**1982**, 6, 403–412. [Google Scholar] [CrossRef] [PubMed] - Sadikoglu, F.; Kavalcioglu, C.; Dagman, B. Electromyogram (EMG) signal detection, classification of EMG signals and diagnosis of neuropathy muscle disease. Procedia Comput. Sci.
**2017**, 120, 422–429. [Google Scholar] [CrossRef] - Benatti, S.; Rovere, G.; Bösser, J.; Montagna, F.; Farella, E.; Glaser, H.; Benini, L. A sub-10mW real-time implementation for EMG hand gesture recognition based on a multi-core biomedical SoC. In Proceedings of the 2017 7th IEEE International Workshop on Advances in Sensors and Interfaces (IWASI), Vieste, Italy, 15–16 June 2017; pp. 139–144. [Google Scholar]
- Vitale, A.; Donati, E.; Germann, R.; Magno, M. Neuromorphic Edge Computing for biomedical applications: Gesture Classification using EMG Signals. IEEE Sens. J.
**2022**, 22, 19490–19499. [Google Scholar] [CrossRef] - Bottou, L.; Bengio, Y. Convergence properties of the K-means algorithms. Adv. Neural Inf. Process. Syst.
**1994**, 7, 585. [Google Scholar] - Kodinariya, T.M.; Makwana, P.R. Review on determining number of Cluster in K-Means Clustering. Int. J.
**2013**, 1, 90–95. [Google Scholar] - Quackenbush, J. Computational analysis of microarray data. Nat. Rev. Genet.
**2001**, 2, 418–427. [Google Scholar] [CrossRef] [PubMed] - Alimisis, V.; Gennis, G.; Touloupas, K.; Dimas, C.; Gourdouparis, M.; Sotiriadis, P.P. Gaussian Mixture Model classifier analog integrated low-power implementation with applications in fault management detection. Microelectron. J.
**2022**, 126, 105510. [Google Scholar] [CrossRef] - Delbrueck, T.; Mead, C. Bump circuits. In Proceedings of the International Joint Conference on Neural Networks, Nagoya, Japan, 25–29 October 1993; Volume 1, pp. 475–479. [Google Scholar]
- Gourdouparis, M.; Alimisis, V.; Dimas, C.; Sotiriadis, P.P. An ultra-low power,±0.3 V supply, fully-tunable Gaussian function circuit architecture for radial-basis functions analog hardware implementation. AEU-Int. J. Electron. Commun.
**2021**, 136, 153755. [Google Scholar] [CrossRef] - Alimisis, V.; Gourdouparis, M.; Dimas, C.; Sotiriadis, P.P. A 0.6 V, 3.3 nW, Adjustable Gaussian Circuit for Tunable Kernel Functions. In Proceedings of the 2021 34th SBC/SBMicro/IEEE/ACM Symposium on Integrated Circuits and Systems Design (SBCCI), Campinas, Brazil, 23–27 August 2021; pp. 1–6. [Google Scholar]
- Villamizar, D.A.; Muratore, D.G.; Wieser, J.B.; Murmann, B. An 800 nW Switched-Capacitor Feature Extraction Filterbank for Sound Classification. IEEE Trans. Circuits Syst. I Regul. Pap.
**2021**, 68, 1578–1588. [Google Scholar] [CrossRef] - Zhang, Y.; Mirchandani, N.; Onabajo, M.; Shrivastava, A. RSSI Amplifier Design for a Feature Extraction Technique to Detect Seizures with Analog Computing. In Proceedings of the 2020 IEEE International Symposium on Circuits and Systems (ISCAS), Seville, Spain, 12–14 October 2020; pp. 1–5. [Google Scholar]
- Yang, M.; Liu, H.; Shan, W.; Zhang, J.; Kiselev, I.; Kim, S.J.; Enz, C.; Seok, M. Nanowatt acoustic inference sensing exploiting nonlinear analog feature extraction. IEEE J. Solid-State Circuits
**2021**, 56, 3123–3133. [Google Scholar] [CrossRef] - Yoo, J.; Yan, L.; El-Damak, D.; Altaf, M.A.B.; Shoeb, A.H.; Chandrakasan, A.P. An 8-cannel scalable EEG acquisition SoC with patient-specific seizure classification and recording processor. IEEE J. Solid-State Circuits
**2012**, 48, 214–228. [Google Scholar] [CrossRef] - Sharma, A.K.; Madhusudan, M.; Burns, S.M.; Mukherjee, P.; Yaldiz, S.; Harjani, R.; Sapatnekar, S.S. Common-centroid layouts for analog circuits: Advantages and limitations. In Proceedings of the IEEE 2021 Design, Automation & Test in Europe Conference & Exhibition (DATE), Grenoble, France, 1–5 February 2021; pp. 1224–1229. [Google Scholar]
- Kang, K.; Shibata, T. An on-chip-trainable Gaussian-kernel analog support vector machine. IEEE Trans. Circuits Syst. I Regul. Pap.
**2009**, 57, 1513–1524. [Google Scholar] [CrossRef] - Chakrabartty, S.; Cauwenberghs, G. Sub-microwatt analog VLSI trainable pattern classifier. IEEE J. Solid-State Circuits
**2007**, 42, 1169–1179. [Google Scholar] [CrossRef][Green Version] - Genov, R.; Cauwenberghs, G. Kerneltron: Support vector “machine” in silicon. IEEE Trans. Neural Netw.
**2003**, 14, 1426–1434. [Google Scholar] [CrossRef][Green Version] - Peng, S.Y.; Hasler, P.E.; Anderson, D. An analog programmable multi-dimensional radial basis function based classifier. In Proceedings of the IEEE 2007 IFIP International Conference on Very Large Scale Integration, Atlanta, GA, USA, 15–17 October 2007; pp. 13–18. [Google Scholar]
- Alimisis, V.; Gennis, G.; Dimas, C.; Gourdouparis, M.; Sotiriadis, P.P. An ultra low power analog integrated radial basis function classifier for smart IoT systems. Analog. Integr. Circuits Signal Process.
**2022**, 1–12. [Google Scholar] [CrossRef] - Alimisis, V.; Gennis, G.; Dimas, C.; Sotiriadis, P.P. An Analog Bayesian Classifier Implementation, for Thyroid Disease Detection, based on a Low-Power, Current-Mode Gaussian Function Circuit. In Proceedings of the IEEE 2021 International Conference on Microelectronics (ICM), New Cairo City, Egypt, 19–22 December 2021; pp. 153–156. [Google Scholar]
- Zhang, R.; Shibata, T. An analog on-line-learning K-means processor employing fully parallel self-converging circuitry. Analog Integr. Circuits Signal Process.
**2013**, 75, 267–277. [Google Scholar] [CrossRef]

**Figure 1.**The utilized bump circuit. The parameter voltages ${V}_{r}$, ${V}_{c}$, and the bias current ${I}_{bias}$ control the Gaussian function’s mean value, variance, and peak value, respectively.

**Figure 2.**The output current of the modified (orange) and the baseline (blue) bump circuits. The controlling parameters for both circuits are set to ${V}_{r}=0$ V, ${V}_{c}=150$ mV, and ${I}_{b}ias=6$ nA.

**Figure 3.**By connecting 10 simple bump circuits sequentially, the output of the last one is equivalent to a 10-D Gaussian function. Each bump circuit’s parameters (${V}_{r}$, ${V}_{c}$, ${I}_{bias}$) are tuned independently.

**Figure 4.**

**Left**: A N neuron NMOS-based Lazzaro WTA circuit.

**Right**: a modified NMOS-based Lazzaro WTA neuron. The PMOS-based variants for both circuits are built accordingly.

**Figure 5.**Comparison between the baseline cascaded WTA and the proposed cascaded WTA circuits with the added diode-connected transistor.

**Figure 6.**The utilized WTA circuit is composed of a 19-neuron WTA and a 4-neuron WTA circuit connected sequentially. The output currents of the first WTA circuit that correspond to the sub-classes of each class are summed and then inserted into the second WTA.

**Figure 7.**The implementation of the analog centroid-based classifier. The output currents of each multivariate bump circuit representing each sub-class are inserted into the WTA circuit to indicate the winning class.

**Figure 8.**The system-level implementation of the proposed voting classifier is composed of 3 processing parts: (

**center**) analog feature extraction (FE) (

**middle**), analog classifiers (

**right**), and analog voting circuit.

**Figure 9.**Top-level concept architecture for a fully analog hand gesture recognition system. It receives an EMG signal using a biosensor, which is then processed to enable the extraction of the classification features. Lastly, the final decision (in a digital format) is achieved via an analog classifier without the need for typical analog-to-digital converters.

**Figure 10.**Layout of the proposed voting classifier, including the 3 centroid-based classifiers and the voting circuit.

**Figure 12.**Classification accuracy histogram of the analog centroid-based classifier attached to the first electrode (classifier 1) over 20 iterations.

**Figure 13.**Classification accuracy histogram of the analog centroid-based classifier attached to the second electrode (classifier 2), over 20 iterations.

**Figure 14.**Classification accuracy histogram of the analog centroid-based classifier attached to the third electrode (classifier 3) over 20 iterations.

**Figure 16.**Monte Carlo histogram of the analog centroid-based classifier attached to the first electrode for $N=100$ points.

**Table 1.**MOS Transistor Dimensions (Figure 1).

NMOS Differential Block | W/L (${\mathit{\mu}}_{\mathit{m}}/{\mathit{\mu}}_{\mathit{m}}$) | Current Correlator | W/L (${\mathit{\mu}}_{\mathit{m}}/{\mathit{\mu}}_{\mathit{m}}$) |
---|---|---|---|

${M}_{n1}$,${M}_{n4}$ | $2.8/0.4$ | ${M}_{p1}$,${M}_{p2}$ | $1.6/1.6$ |

${M}_{n2}$,${M}_{n3}$ | $0.4/0.4$ | ${M}_{p3}$-${M}_{p6}$ | $0.4/1.6$ |

${M}_{n5}$-${M}_{n8}$ | $0.4/1.6$ | - | - |

${M}_{n9}$,${M}_{n10}$ | $1.6/1.6$ | - | - |

Method | Best | Worst | Mean | Std. |
---|---|---|---|---|

Software Voting classifier | $0.954$ | $0.915$ | $0.938$ | $0.012$ |

Analog Voting classifier | $0.932$ | $0.893$ | $0.912$ | $0.011$ |

Software Classifier 1 | $0.904$ | $0.871$ | $0.887$ | $0.010$ |

Analog Classifier 1 | $0.904$ | $0.85$ | $0.875$ | $0.014$ |

Software Classifier 2 | $0.916$ | $0.858$ | $0.889$ | $0.016$ |

Analog Classifier 2 | $0.906$ | $0.853$ | $0.886$ | $0.014$ |

Software Classifier 3 | $0.845$ | $0.797$ | $0.825$ | $0.014$ |

Analog Classifier 3 | $0.844$ | $0.777$ | $0.805$ | $0.017$ |

Method | Best | Worst | Mean | Std. |
---|---|---|---|---|

Analog Voting classifier | $0.915$ | $0.901$ | $0.912$ | $0.004$ |

Analog Classifier 1 | $0.910$ | $0.565$ | $0.884$ | $0.038$ |

Technology | Classifier | No. of Dimensions | Power Consumption | Energy Per Classification | Area | |
---|---|---|---|---|---|---|

This Work | 90 nm | Voting | 10 | 31.5 $\mathsf{\mu}$W | $\frac{225\phantom{\rule{3.33333pt}{0ex}}\mathrm{pJ}}{\mathrm{classification}}$ | 1.67 mm${}^{2}$ |

[37] | 90 nm | GMM | 16 | 12.0 $\mathsf{\mu}$W | $\frac{96\phantom{\rule{3.33333pt}{0ex}}\mathrm{pJ}}{\mathrm{classification}}$ | 0.451 mm${}^{2}$ |

[46] | 0.18 $\mathsf{\mu}$m | SVM | 2 | 220.0 $\mathsf{\mu}$W | $\frac{252\phantom{\rule{3.33333pt}{0ex}}\mathrm{pJ}}{\mathrm{vector}}$ | 0.060 mm${}^{2}$ |

[47] | 0.5 $\mathsf{\mu}$m | SVM | 14 | 840.0 nW | $\frac{21\phantom{\rule{3.33333pt}{0ex}}\mathrm{nJ}}{\mathrm{classification}}$ | 9.000 mm${}^{2}$ |

[48] | 0.5 $\mathsf{\mu}$m | SVM | N/A | 5.9 mW | $\frac{460\phantom{\rule{3.33333pt}{0ex}}\mathrm{pJ}}{\mathrm{sample}}$ | 9.000 mm${}^{2}$ |

[49] | 0.5 $\mathsf{\mu}$m | RBF NN | 2 | N/A | N/A | 2.250 mm${}^{2}$ |

[50] | 90 nm | GRBFN | 7 | 330 nW | $\frac{2\phantom{\rule{3.33333pt}{0ex}}\mathrm{pJ}}{\mathrm{vector}}$ | 0.050 mm${}^{2}$ |

[51] | 90 nm | Bayesian | 5 | 365 nW | $\frac{2.15\phantom{\rule{3.33333pt}{0ex}}\mathrm{pJ}}{\mathrm{classification}}$ | 0.030 mm${}^{2}$ |

[52] | 0.18 $\mathsf{\mu}$m | K-means | 164 | N/A | N/A | N/A |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Alimisis, V.; Mouzakis, V.; Gennis, G.; Tsouvalas, E.; Dimas, C.; Sotiriadis, P.P. A Hand Gesture Recognition Circuit Utilizing an Analog Voting Classifier. *Electronics* **2022**, *11*, 3915.
https://doi.org/10.3390/electronics11233915

**AMA Style**

Alimisis V, Mouzakis V, Gennis G, Tsouvalas E, Dimas C, Sotiriadis PP. A Hand Gesture Recognition Circuit Utilizing an Analog Voting Classifier. *Electronics*. 2022; 11(23):3915.
https://doi.org/10.3390/electronics11233915

**Chicago/Turabian Style**

Alimisis, Vassilis, Vassilis Mouzakis, Georgios Gennis, Errikos Tsouvalas, Christos Dimas, and Paul P. Sotiriadis. 2022. "A Hand Gesture Recognition Circuit Utilizing an Analog Voting Classifier" *Electronics* 11, no. 23: 3915.
https://doi.org/10.3390/electronics11233915