Investigate, and train, for Armenian and Russian the following sentence segmentation tools:

http://wiki.apertium.org/wiki/Sentence_tokenisation

For each tool there should be a HOWTO for training.

You may need to email the original authors.