显示 python opencv 中的视觉词袋中的视觉词示例/样本

Question

I have been working with the Bag of visual words method on my project and would like to produce a visual example of my image words.

My Bag of Visual words method, coded in python on opencv3, operates correctly, creating the vocabulary and the image vectors as required.

However I'd like to be able to have some sort of visual output for the clusters, much like this example:

What would be a way to achieve this? I have tried feature matching utilizing the vocabulary versus a singular image form my data sets descriptors but since the image word is a descriptor not a key point, this method does not work out.

Any help would be greatly appreciated.

Answer 1

Are you trying to just get the data displayed that way? If so, the fastest way would be to look through the AI explainability libraries already created:

https://github.com/EthicalML/awesome-production-machine-learning#explaining-black-box-models-and-datasets

If you want to build that, note that the individual neurons in most neural networks (variational autoencoders are an exception) don't easily produce images that look as nice as the ones in those images, so you will want to look at the research papers listed in the explainability projects, like this one: https://github.com/yosinski/deep-visualization-toolbox

However, if you do have convolutions with tensors that look exactly like that when displayed, then this is a problem of organizing and displaying those in some GUI, like pyQT or kivy.