1) The algorithm computes transformation invariant keypoints represented as a binary vector of [-1,1] , meaning the feature vector contains the same information for all projections from the same molecule(https://ieeexplore.ieee.org/abstract/document/9169844).
1) The algorithm encodes a image into a feature vector.
2) The vector expresses key characteristics of the protein, each pixel of the computed 16 x 16 image is weighed by a value between [0,max]. 4 areas of the protein are is extracted used for training of the neural network, improving the classification. ()
2) the generator utilizes the feature vector to recreate the 3D model and reproject it onto the 2D plane and apply a known CTF and a trainable Noise.
3) Each protein component is represented as a binarized vector which is concatenated with the other part , partial and full image vectors, improving the overall accuracy(https://arxiv.org/pdf/1902.09941.pdf).
3) The discriminator performs a featurewise loss to check how real the reproduced model is compared to the data. We then cluster the feature vectos.