Methods, systems, and techniques for 3D or 2.5D electronic communication, such as holographic communication. Two-dimensional image data is generated by a camera that has imaged at least part of a three-dimensional head including a face of a conference participant is obtained. A photo-realistic 2.5-dimensional or three-dimensional representation of at least part of the head is reconstructed using the two-dimensional image data. The three-dimensional representation of the at least part of the head includes a reconstruction of an area missing from the two-dimensional image data. One or more artificial neural networks, such as but not necessarily limited to a convolutional neural network and/or a multilayer perceptron neural network, may be used for the reconstruction.