TransJury: Towards explainable transfer learning through selection of layers from deep neural networks