Abstract: The main theme of this project is to Enhance and Color Correction for underwater images. This project becomes challenging due to attenuation and scattering of light. In this process, the novel algorithm of deep learning algorithms along with gamma correction. In the procedure of enhancing the texture and structural preservation is more important. In this work, the image enhancement is obtained by using the convolution neural networks. This process involves two stages mainly the training and testing stage. During training process, the dataset is collected, and their up sampled and resized images are stored in a mat file. Then CNN layers are created. Finally, the train the network using the data stored in mat files and CNN layers. After the training process, the test image is given as input to network designed earlier. Then finally the high-resolution image is obtained. This method reduces the loss of textural and structural information when compared to state of art methods.
Keywords: CNN, Deep Learning, Underwater Images.
Cite:
Varun N, Mrs Shaila V Hegde,"Enhancement For Underwater Images Using Hybrid Deep Learning", IJARCCE International Journal of Advanced Research in Computer and Communication Engineering, vol. 12, no. 11, pp. 131-141, 2023, Crossref https://doi.org/10.17148/IJARCCE.2023.121119.
|
DOI:
10.17148/IJARCCE.2023.121119
