1 of 4

Machine Vision Technology October 2023

Assignment
Machine Vision Toward Human Activity
Recognition In Rehabilitation
Faris Idris Bin Aminuddin – 50226121251
Mechatronics Engineering Technology (MECHA), Industrial Automation Section, UniKL MFI
[email protected]

Introduction
Innovative study methods and technological advances are causing rehabilitation
research to change quickly. Finding out how intrinsic drive changes during cognitive
rehabilitation is the first step towards finding successful ways to combine approaches in
EEG-based BCI rehabilitation. [1] , [2] Machine learning methods are being tested to see
how well they work at figuring out the best exercises for people with low back pain.
These improvements are like those seen in location and tracking modes for sports
rehabilitation training. [3] , [4]

The narrative is changing because of a look at interactive computing technologies for

improving the lives of stroke survivors and the use of machine learning in wearable
devices that recognize frozen shoulder exercises. [5] , [6] These chosen studies give us
information about a field that is about to change the way therapy is done by using
cutting-edge tools and all-encompassing methods.

Body of knowledge 1
Assessments of intrinsic motivation during cognitive rehabilitation must be
sophisticated to understand motivation dynamics. [1] Innovative methods are used to
rethink intrinsic motivation in cognitive recovery. The merging of connection network
characteristics with temporal-spatial analysis revolutionizes EEG-based BCI
rehabilitation. [2] These patterns improve brain-computer interface dynamics
knowledge, offering dramatic rehabilitative outcomes. In low back pain physical
rehabilitation exercises, machine learning algorithms evaluate efficiency and
performance. [3] This study assesses algorithmic efficacy and optimizes these
activities, enabling more personalized rehabilitation programmers.

Body of knowledge 2
A new paradigm in sports rehabilitation training is revealed by self-powered sensors
and particle swarm optimization algorithms. [4] This dynamic method tracks location
and mode, revolutionizing athlete rehabilitation and promising improved training
precision. For post-stroke therapy, highly interactive and immersive computing
technologies open new possibilities. [5] This study showcases post-stroke healing
methods of the future, from virtual to real-world. Frozen shoulder rehabilitation using
machine learning and wearable recognition is crucial. [6] This technology-rehabilitation
combination will revolutionize frozen shoulder recovery and management with
personalized treatment.
 2 of 4

Body of knowledge 3
Human activity recognition with wearable sensors and a deep CNN-LSTM model is
revolutionary. [7] This combination records human actions and advances assistive
technologies, especially for monitoring and analyzing daily movements. A mixed deep
learning model for geriatric activity recognition and fall detection advances eldercare.
[8]
This novel model explores the difficulties of senior daily life and shows how
technology will become an integral aspect of aged care. Rehabilitation technologies are
entering a new era with the preliminary assessment of a postural synergy-based
exoskeleton for post-stroke upper limb rehabilitation. [9] This research examines the
viability of such devices and suggests a more personalized and successful post-stroke
therapy. Understanding the characteristics that influence US post-stroke sensorimotor
rehabilitation device adoption. [10] provides crucial insights. This study explores device
adoption patterns and provides a path for integrating novel rehabilitation technology
into mainstream healthcare.

Conclusion
To summarize, the combination of many rehabilitation methods, ranging from cognitive
to physical aspects, demonstrates the ever-evolving field of modern rehabilitation
research. The use of sophisticated technology, such as EEG-based brain-computer
interfaces (BCI) and machine learning algorithms, signifies a fundamental change in
approach towards more efficient and tailored rehabilitation methods. These
developments have the potential to fundamentally change the way rehabilitation is
approached and carried out, leading to better results, and improving the overall quality
of life for persons undergoing different rehabilitation processes. [1] , [2] , [3]

Table 1. Summary of literature

Ref. Yea Description Activity Sensor/ Classifier Result/ Dataset
r Data Performanc
e
[1] 2021 Novel method Cognitive EEG, Innovative Redefining Patient-specific
evaluating changes in activities Behavioral assessment intrinsic cognitive
intrinsic motivation data) methodologies motivation rehabilitation data
during cognitive understanding
rehabilitation.
[2] 2022 Effective fusing Brain- EEG data, Connectivity Enhanced BCI BCI dataset,
approach for EEG-based computer Connectivity network rehabilitation Connectivity
BCI rehabilitation. interface network patterns, outcomes patterns
patterns Temporal-
spatial analysis
[3] 2023 Analyzing data efficiency Low back Sensor data, Various ML Optimization Low back pain
and performance of ML pain Machine algorithms potential, exercise dataset
algorithms for low back exercises learning Efficiency
pain rehabilitation. algorithms evaluation
[4] 2023 Location and tracking Sports Self-powered Particle swarm Enhanced Sports training
mode of sports rehabilitation sensors, optimization training dataset
rehabilitation training. training Particle swarm algorithm precision,
optimization Dynamic
algorithm tracking
[5] 2023 Review of application of Post-stroke Interactive Immersive Revolutionary Post-stroke
highly interactive and survivor's computing computing impact, Future survivor's virtual
immersive computing rehabilitation technologies technologies possibilities environment data
technologies for post-
stroke survivor's
rehabilitation.
 3 of 4

[6] 2023 Wearable-based frozen Frozen Wearable Machine Redefining Frozen shoulder
shoulder rehabilitation shoulder sensors, learning healthcare rehabilitation
exercise recognition rehabilitation Machine approaches approach, dataset
using ML approaches. learning Personalized
approaches care

[7] 2022 Deep CNN-LSTM with Human Wearable Deep CNN- Transformative Human activity
self-attention model for activity sensors, Deep LSTM model paradigm, dataset
human activity recognition CNN-LSTM with self- Assistive
recognition using model attention technologies
wearable sensor.
[8] 2023 Hybrid deep learning Elderly care Wearable Hybrid deep Advancements Elderly activity
model for human sensors, learning in eldercare, dataset
activity recognition and Hybrid deep model Complex daily
fall detection for the learning model activities
elderly.
[9] 2021 Preliminary assessment Post-stroke Postural Feasibility Personalized Post-stroke upper
of a postural synergy- upper limb synergy-based assessment rehabilitation limb rehabilitation
based exoskeleton for rehabilitation exoskeleton devices dataset
post-stroke upper limb
rehabilitation.
[10] 2021 Determining factors that Sensorimotor Adoption Market Critical Sensorimotor
influence adoption of rehabilitation factors adoption insights, rehabilitation
new post-stroke devices dynamics Integration adoption dataset
sensorimotor roadmap
rehabilitation devices in
the USA.

References

[1] Y. Nishiwaki, M. Nakamura, and M. Nihei, “A novel method of evaluating changes in intrinsic motivation during
cognitive rehabilitation,” in Proceedings of the Annual International Conference of the IEEE Engineering in Medicine
and Biology Society, EMBS, Institute of Electrical and Electronics Engineers Inc., 2021, pp. 2095–2099. doi:
10.1109/EMBC46164.2021.9630738.
[2] L. Cao et al., “An Effective Fusing Approach by Combining Connectivity Network Pattern and Temporal-spatial Analysis
for EEG Based BCI Rehabilitation,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2022, doi:
10.1109/TNSRE.2022.3198434.
[3] A. Marusic, L. Annabi, S. M. Nguyen, and A. Tapus, “Analyzing Data Efficiency and Performance of Machine Learning
Algorithms for Assessing Low Back Pain Physical Rehabilitation Exercises,” in Proceedings of the 11th European
Conference on Mobile Robots, ECMR 2023, Institute of Electrical and Electronics Engineers Inc., 2023. doi:
10.1109/ECMR59166.2023.10256318.
[4] Y. Ke, “Location and Tracking Mode of Sports Rehabilitation Training With Self-Powered Sensors Based on Particle
Swarm Optimization Algorithm,” IEEE Sens J, vol. 23, no. 18, pp. 20894–20903, Sep. 2023, doi:
10.1109/JSEN.2023.3241269.
[5] S. Shankar, S. J. Syed Ali Fathima, M. Uma Priya, S. Shriram, S. Priyadarshini, and A. Benazir Begum, “Review of
Application of Highly Interactive and Immersive Computing Technologies for Enhancement of Post-Stroke Survivor’s
Rehabilitation,” in 2nd International Conference on Sustainable Computing and Data Communication Systems, ICSCDS
2023 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 2023, pp. 647–654. doi:
10.1109/ICSCDS56580.2023.10104892.
[6] C. P. Liu, C. C. Lai, K. C. Liu, C. Y. Hsieh, and C. T. Chan, “Wearable-based Frozen Shoulder Rehabilitation Exercise
Recognition using Machine Learning Approaches,” in 2023 IEEE International Symposium on Medical Measurements
and Applications, MeMeA 2023 - Conference Proceedings, Institute of Electrical and Electronics Engineers Inc., 2023.
doi: 10.1109/MeMeA57477.2023.10171854.
[7] M. A. Khatun et al., “Deep CNN-LSTM With Self-Attention Model for Human Activity Recognition Using Wearable
Sensor,” IEEE J Transl Eng Health Med, vol. 10, 2022, doi: 10.1109/JTEHM.2022.3177710.
[8] F. Kharrat, W. Gueaieb, F. Karray, and A. Elsaddik, “A Hybrid Deep Learning Model for Human Activity Recognition and
Fall Detection for the Elderly,” in 2023 IEEE International Symposium on Medical Measurements and Applications,
MeMeA 2023 - Conference Proceedings, Institute of Electrical and Electronics Engineers Inc., 2023. doi:
10.1109/MeMeA57477.2023.10171926.
[9] C. He, C. H. Xiong, Z. J. Chen, W. Fan, X. L. Huang, and C. Fu, “Preliminary Assessment of a Postural Synergy-Based
Exoskeleton for Post-Stroke Upper Limb Rehabilitation,” IEEE Transactions on Neural Systems and Rehabilitation
 4 of 4
Engineering, vol. 29, pp. 1795–1805, 2021, doi: 10.1109/TNSRE.2021.3107376.
[10] C. M. Morrow, E. Johnson, K. N. Simpson, and N. J. Seo, “Determining factors that influence adoption of new post-
stroke sensorimotor rehabilitation devices in the USA,” IEEE Transactions on Neural Systems and Rehabilitation
Engineering, vol. 29, pp. 1213–1222, 2021, doi: 10.1109/TNSRE.2021.3090571.