* refine README * Update README.md
* init README of deepdoc, add picture processer. * add resume parsing