Show error message, etc.

[yasunorikudo]

• Add a description about a recommended opencv environment.
• Replace the trained model with more accurate one.
• Show an error message if enough key points were not detected. For example,

$ python bin/demo.py sample/image_wrong.png --lift_model sample/gen_epoch_6.npz --model2d pose_iter_440000.caffemodel --proto2d openpose_pose_coco.prototxt
MODEL: generator, N_OUT: 17, N_UNIT: 1024
Traceback (most recent call last):
  File "bin/demo.py", line 205, in <module>
    main(args)
  File "bin/demo.py", line 168, in main
    '{}/{} points were detected. Try another image.'.format(n_detect_key_points, len(points)))
Exception: OpenPose could not detect enough key points. 13/19 points were detected. Try another image.

[kogaki]

Great! This must care the issue #2 !

[yasunorikudo]

Yes, this PR solves the issues #2 and #3.

[yasunorikudo]

I noticed that detecting 19 key points is not necessary to convert to Human3.6M’s key points. So I will fix.