Can CMUSphinx make distinctions between speech and background noise/music

Question

I've been fiddling with Sphinx's word searching functionality for a bit now and have come to the conclusion that sorting through music to find words is a lost cause. However I don't need Sphinx to make such rigid distinctions. I only want to know where a potential word begins and ends, or more preferably where a phrase ends and one begins. Is there any way I can get a probabilistic reading on sounds to see a sound(s) potential for being word.

import os, sys
from pocketsphinx import AudioFile

fps=100
config = {
    'verbose': False,
    'audio_file': sys.argv[2],
    'buffer_size': 2000,
    'frate': fps,
    'no_search': False,
    'keyphrase': sys.argv[1],
    'hmm': '../../git/pocketsphinx/model/en-us/en-us',
    'lm': False,
    'dict': '../../git/pocketsphinx/model/en-us/cmudict-en-us.dict'
}

audio = AudioFile(**config)
for phrase in audio:
    for s in phrase.seg():
        print('| %4ss | %4ss | %8s |' % (s.start_frame / fps, s.end_frame / fps, s.word)

Could I make any modifications to the above code to get this sort of data? Thanks in advance.