Image Colorization
Main purpose of doing this is to revive the historical images/ art galleries and give them a new meaning via realistic colorization. What is missing in our current knowledge is the quality of the camera in smartphone used to scan the image as well the environment where it is being scanned dep
2025-06-28 16:33:01 - Adil Khan
Image Colorization
Project Area of Specialization Artificial IntelligenceProject SummaryMain purpose of doing this is to revive the historical images/ art galleries and give them a new meaning via realistic colorization.
What is missing in our current knowledge is the quality of the camera in smartphone used to scan the image as well the environment where it is being scanned depending on the light exposure.
The Main Features of this application will be:
- Deep learning.
- Neural networks.
- Image Processing.
- Mobile/Web Application.
- Cloud Computing.
The project will meet the following objectives:
- Scan a standard pixel size old grayscale photographs and identity its aspects.
- Convert the scanned image into colored image.
- Mobile/Web Based Application to upload image.
- Using Deep Learning our model will be trained by using several different images.
- All the processing will be done on Cloud while the colored image will be sent on Mobile/Web App.
- A set of historical images will be colorized.
- Image processing using Convolutional Neural Network to change the image into colored image
Our approach is based on Convolutional Neural Networks that will learn complex mappings and will be able to predict colors using several colored sample images. Neural network is formed by several further subcomponents.
A fully automatic approach produces vibrant and realistic colorizations using Deep Learning techniques and methods.
We will train our system to identify basic colors so that it can convert greyscale / dim / yellowish color pictures to realistic colorizations.
Main purpose of doing this is to revive the historical images/ art galleries and give them a new meaning via realistic colorization.
Technical Details of Final Deliverable- Project Deliverable 1: Scan up to 1200 x 1600 pixel size old grayscale photographs and identity its aspects.
- Project Deliverable 2: Convert the scanned image into colored image.
- Product Deliverable 1: Mobile/Web Based Application to upload image.
- Product Deliverable 2: Automated Image Colorization without any user interference
- Product Deliverable 3: All Process to be done on cloud to make it compatible for low end devices