Đồ án tốt nghiệp - Development and optimization of a custom multi-task cascaded cnn for EEFD

HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND DUCATION
FACULTY OF INTERNATIONAL EDUCATION

GRADUATION PROJECT

DEVELOPMENT AND OPTIMIZATION OF A
CUSTOM MULTI-TASK CASCADED CNN FOR
EFFICIENT EDGE FACE DETECTION

PHAN MINH DUC
Student ID: 21119303
NGUYEN LE NGOC TRAM
Student ID: 21119322
Major: COMPUTER ENGINEERING TECHNOLOGY
Advisor: PHAM VAN KHOA, PhD.

Ho Chi Minh City, July 2025

HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND DUCATION
FACULTY OF INTERNATIONAL EDUCATION

GRADUATION PROJECT

DEVELOPMENT AND OPTIMIZATION OF A
CUSTOM MULTI-TASK CASCADED CNN FOR
EFFICIENT EDGE FACE DETECTION

PHAN MINH DUC
Student ID: 21119303
NGUYEN LE NGOC TRAM
Student ID: 21119322
Major: COMPUTER ENGINEERING TECHNOLOGY
Advisor: PHAM VAN KHOA, PhD.

Ho Chi Minh City, July 2025

DISCLAIMER
This graduation thesis was completed as part of the requirements for the bachelor’s degree
in computer engineering technology at Ho Chi Minh City University of Technology and
Education. The content presented in this thesis is the result of our independent research and
development on the topic “Development and Optimization of a Custom Multi-task
Cascaded CNN for Efficient Edge Face Detection”, conducted under the supervision of
Pham Van Khoa, PhD.
We have made every effort to ensure the accuracy and completeness of the technical
information, experimental results, and analyses included herein. All data sources,
algorithms, frameworks, and scientific references utilized in this work have been properly
cited and acknowledged in accordance with academic standards and ethical research
practices.
The findings, interpretations, and conclusions expressed in this thesis represent the views
of the authors and do not necessarily reflect the official policy, position, or opinions of Ho
Chi Minh City University of Technology and Education, its faculty, or any affiliated
organizations. This work is intended solely for academic and educational purposes and
does not constitute professional or industrial advice.
The authors do not guarantee the completeness, precision, reliability, suitability, or
availability of the methodologies, software implementations, hardware configurations, or
visual materials presented in this thesis for any specific application. Any use or reliance on
the content of this thesis is strictly at the user’s own risk. The university and supervising
faculty disclaim responsibility for any errors, omissions, or consequences arising from the
application of this re
...

...
rgess, and Natasha Noy. Google Dataset Search: Building
a search engine for datasets in an open Web ecosystem. The world wide web conference.
2019.
[12] Liu, Bing, et al. An FPGA-based CNN accelerator integrating depthwise separable
convolution. Electronics 8.3 (2019): 281.
[13] Bai, Lin, Yiming Zhao, and Xinming Huang. A CNN accelerator on FPGA using
depthwise separable convolution. IEEE Transactions on Circuits and Systems II: Express
Briefs 65.10 (2018): 1415-1419.
76

[14] Chollet, François. Xception: Deep learning with depthwise separable convolutions.
Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.
[15] Liu, Gaiping, Jianmei Xiao, and Xihuai Wang. Optimization of Face Detection
Algorithm based on MTCNN. International Core Journal of Engineering 7.8 (2021): 456464.
[16] Ma, Long‐Hua, et al. Acceleration of multi‐task cascaded convolutional networks.
IET Image Processing 14.11 (2020): 2435-2441.
[17] Howard, Andrew G., et al. Mobilenets: Efficient convolutional neural networks for
mobile vision applications. arXiv preprint arXiv:1704.04861 (2017).
[18] Hu, Jie, Li Shen, and Gang Sun. Squeeze-and-excitation networks. Proceedings of
the IEEE conference on computer vision and pattern recognition. 2018.
[19] Jin, Xin, et al. Delving deep into spatial pooling for squeeze-and-excitation
networks. Pattern Recognition 121 (2022): 108159.
[20] Ge, Rongjun, et al. Convolutional squeeze-and-excitation network for ECG
arrhythmia detection. Artificial Intelligence in Medicine 121 (2021): 102181.
[21] Rokh, Babak, Ali Azarpeyvand, and Alireza Khanteymoori. A comprehensive survey
on model quantization for deep neural networks in image classification. ACM
Transactions on Intelligent Systems and Technology 14.6 (2023): 1-50.
[22] Cheng, Jian, et al. Quantized CNN: A unified approach to accelerate and compress
convolutional networks. IEEE transactions on neural networks and learning
systems 29.10 (2017): 4730-4743.
[23] Sailesh, M., and K. Selvakumar. A novel framework for deployment of CNN models
using post-training quantization on microcontroller. Microprocessors and
Microsystems 94 (2022): 104634.
[24] Kluska, Piotr, and Maciej Zięba. Post-training quantization methods for deep learning
models. Intelligent Information and Database Systems: 12th Asian Conference, ACIIDS
2020, Phuket, Thailand, March 23–26, 2020, Proceedings, Part I 12. Springer
International Publishing, 2020.

77