All projects03 / 16
Super Image Quality Enhancer (SIQE)
Enhances low-resolution images using AI focused on the luminance channel.
Developing a deep learning model utilizing residual dense blocks (RDBs) to enhance image quality by focusing on its luminance (Y) component
The problem
Low-resolution images lose detail when enlarged using conventional resizing. SIQE addresses this with a learned super-resolution pipeline and a web interface for comparing the original image with its enhanced output.
The approach
- 01
Prepare image datasets and extract the luminance channel for training.
- 02
Configure and train a model built around Residual Dense Blocks.
- 03
Load the trained model through the FastAPI backend.
- 04
Process images through the enhancement pipeline.
- 05
1. Compare original and enhanced images in the web interface.