Images are composed of varying hues of color. Digitally, the composition of RGB pixels is what is projected from the screens of our devices to display an image. Transferring RGB pixels from one image to another is a rather interesting question to ask: what does a flower look like when it takes the color of a mountain scenery?
In this workshop, we will look at how to use clustering to reduce tens of thousands of unique RGB pixels down to tens of the most important pixels, a technique used in the early days of gaming. Also, we will look at transferring pixels in reduced color space, understand the complexity of the seemingly insurmountable problem of transferring whole palettes, and find out why the RGB color space is “bad”. We’ll use NumPy, scikit-learn, and Dask via Colab for this purpose.
This session will be conducted by Syafiq Kamarul Azman, research engineer at Khalifa University and a self-professed "machine learning geek".
Palette Transfer in Python
Images are composed of varying hues of color. Digitally, the composition of RGB pixels is what is projected from the screens of our devices to display an image. Transferring RGB pixels from one image to another is a rather interesting question to ask: what does a flower look like when it takes the color of a mountain scenery?
In this workshop, we will look at how to use clustering to reduce tens of thousands of unique RGB pixels down to tens of the most important pixels, a technique used in the early days of gaming. Also, we will look at transferring pixels in reduced color space, understand the complexity of the seemingly insurmountable problem of transferring whole palettes, and find out why the RGB color space is “bad”. We’ll use NumPy, scikit-learn, and Dask via Colab for this purpose.
This session will be conducted by Syafiq Kamarul Azman, research engineer at Khalifa University and a self-professed "machine learning geek".
Prerequisites:
— Basic Python knowledge
— Visual Studio Code