Fixing a blurred photograph: blind image deblurring
This project presents a deep learning-based approach to blind image deblurring using a convolutional neural network. The trained model can produce a deblurred output using only the blurred image as input and exhibits improved image quality, as demonstrated by the evaluation of various blurred images...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/166940 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | This project presents a deep learning-based approach to blind image deblurring using a convolutional neural network. The trained model can produce a deblurred output using only the blurred image as input and exhibits improved image quality, as demonstrated by the evaluation of various blurred images. An application has been developed based on this approach in the cloud (Gradio) for image deblurring purposes. Furthermore, we have developed a MATLAB application to compare the effectiveness of the proposed deep learning method with traditional deblurring methods. According to the findings, the deep learning-based deblurring method is a promising solution that provides simplicity, speed, and versatility. Adequate data and training are necessary to enhance the model's capabilities, which may eventually replace traditional deconvolution algorithms in everyday applications. Nevertheless, traditional deconvolution algorithms are still helpful and can provide good results in image deblurring. Hence, we recommend using a hybrid approach that combines both methods for effective image deblurring. |
|---|