Analysis and Research of Image Processing Optimization Technology based on Big Data Processing

Convolutional neural network algorithm is a kind of deep neural network algorithm, which has good performance of image feature extraction. It can reduce the large data amount of images to a small data amount without affecting the feature expression, and effectively solve the problem of low image rec...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	2024 Second International Conference on Data Science and Information System (ICDSIS) s. 1 - 5
Hlavní autoři:	Chen, Yelei, Lan, Ruixue
Médium:	Konferenční příspěvek
Jazyk:	angličtina
Vydáno:	IEEE 17.05.2024
Témata:	Accuracy Big Data big data processing technology Computational modeling convolutional neural network image processing optimization technology Image recognition Mean square error methods PSNR Training
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	Convolutional neural network algorithm is a kind of deep neural network algorithm, which has good performance of image feature extraction. It can reduce the large data amount of images to a small data amount without affecting the feature expression, and effectively solve the problem of low image recognition accuracy. In this paper, the structure of convolutional neural network algorithm is studied deeply, and a 10-layer convolutional neural network model is designed according to the requirements of computer image recognition. Through image transformation and backpropagation algorithm training, the overfitting problem in model training is effectively controlled. The experimental results show that the algorithm has smaller peak signal-to-noise ratio, structural similarity coefficient and mean square error, and the running time is below 0.13s.
DOI:	10.1109/ICDSIS61070.2024.10594155