UppuluriKalyani / ML-Nexus

ML Nexus is an open-source collection of machine learning projects, covering topics like neural networks, computer vision, and NLP. Whether you're a beginner or expert, contribute, collaborate, and grow together in the world of AI. Join us to shape the future of machine learning!
https://discord.gg/n2D4RqnU
MIT License
49 stars 84 forks source link

Image Super -Resolution #483

Open Priyankasanyal04 opened 2 days ago

Priyankasanyal04 commented 2 days ago

Is your feature request related to a problem? Please describe. A clear and concise description of what the problem is. Yes, Image Super-Resolution (ISR) is directly related to solving specific problems in various fields that involve image quality and resolution enhancement. In many cases, images captured or stored at low resolution lack detail, making them less useful or visually appealing. This is common in old photos, compressed images, and low-quality video frames. ISR uses deep learning to enhance the image's resolution, reconstructing finer details, and improving the quality.

Describe the solution you'd like A clear and concise description of what you want to happen. The model takes a low-resolution image as input, typically one that has been downscaled or captured at a lower resolution. The ISR model, often a Super-Resolution Convolutional Neural Network (SRCNN), processes the input image through a series of convolutional layers: First Layer: Extracts low-level features (edges, textures) from the input image. Intermediate Layers: Learn deeper patterns and structures in the image, refining the extracted features. Final Layer: Reconstructs a higher-resolution version of the image using learned features, while preserving important content. The model outputs a high-resolution image that has more detail and clarity than the input. The image is upscaled to the desired resolution. Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

  1. Generative Adversarial Networks (GANs) for ISR
  2. Fast Super-Resolution Models (FSRCNN)
  3. Deep Recursive Residual Network (DRRN)

Approach to be followed (optional) A clear and concise description of the approach to be followed.

Additional context Add any other context or screenshots about the feature request here.

input image output image
github-actions[bot] commented 2 days ago

Thanks for creating the issue in ML-Nexus!🎉 Before you start working on your PR, please make sure to:

github-actions[bot] commented 2 days ago

Thanks for raising this issue! However, we believe a similar issue already exists. Kindly go through all the open issues and ask to be assigned to that issue.

Priyankasanyal04 commented 2 days ago

no such issue exists. @UppuluriKalyani please review it