Author(s)

HaoYang Nie

Affiliation(s)

College of Electronic Engineering, Xi'an Jiaotong Liverpool University, Suzhou 215000, Jiangsu, China.

Corresponding Author

HaoYang Nie

ABSTRACT

Autofocus technology is crucial in many fields, but traditional passive autofocus methods face issues such as low convergence speed, easy misjudgment, and focus breathing. Meanwhile, electronic component detection requires high accuracy and adaptability to practical scenarios. To address these problems, this study constructs an end-to-end electronic component detection baseline and explores the optimization of passive autofocus technology. First, we synthesized images of four electronic components and generated classification datasets as well as multi-object detection datasets. We adopted grayscale downsampling for feature extraction and combined standardization preprocessing with a Support Vector Classifier (SVC) for model training and testing. Additionally, we conducted a comparative analysis between the Convolutional Neural Network (CNN) and Vision Transformer (ViT) models. Experimental results show that the CNN-based detection system has reliable recognition performance for components with distinct morphological features. Compared with ViT, CNN exhibits better adaptability to small datasets, lower computational complexity, and stronger local feature capture capabilities, making it more suitable for practical application scenarios with limited hardware resources. This study provides a feasible baseline for electronic component detection and lays a foundation for the subsequent optimization of passive autofocus technology. 

KEYWORDS

Electronic component detection; Convolutional Neural Network (CNN); Passive autofocus; Vision Transformer (ViT); Model optimization

CITE THIS PAPER

HaoYang Nie. Construction of electronic component detection system based on CNN and optimization of passive autofocus technology. Journal of Trends in Applied Science and Advanced Technologies. 2025, 2(1): 48-52. DOI: https://doi.org/10.61784/asat3016.

REFERENCES

[1] Wang Y, Wu C, Gao Y, et al. Deep learning-based dynamic region of interest autofocus method for grayscale image. Sensors, 2024, 24(13): 4336. DOI: 10.3390/s24134336.

[2] Han J-W, Kim J-H, Lee H-T, et al. A novel training-based auto-focus for mobile-phone cameras. IEEE Transactions on Consumer Electronics, 2011, 57(1): 232-238

[3] Yousefi S, Rahman M, Kehtarnavaz N, et al. A new auto-focus sharpness function for digital and smart-phone cameras. IEEE International Conference on Consumer Electronics (ICCE), 2011: 475-476.

[4] Rahman M T, Kehtarnavaz N. Real-time face-priority auto focus for digital and cell-phone cameras. IEEE Transactions on Consumer Electronics, 2008, 54(4): 1506-1513.

[5] He J, Zhou R, Hong Z. Modified fast climbing search auto-focus algorithm with adaptive step size searching technique for digital camera. IEEE Transactions on Consumer Electronics, 2003, 49(2): 257-262.

[6] Fukushima K. Noncognition: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biological Cybernetics, 1980, 36(4): 193-202.

[7] LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 1998, 86(11): 2278-2324.

[8] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 2012, 25.

[9] Wang Y H, Bai Y, Zhang T. Object detection method based on CNN and camera calibration. Proceedings of the 4th International Conference on Artificial Intelligence, Automation and Algorithms (AI2A '24), 2024: 213-218. DOI: 10.1145/3700523.3700619.

[10] Dosovitskiy A, Beyer L, Kolesnikov A, et al. An image is worth 16×16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929v2 [cs.CV], 2021.