Eu encontrei um programa que faz isso. É pdfsizeopt.py deste local . Eu usei o comando
python pdfsizeopt.py --stats target.pdf
e obteve uma saída como esta:
info: This is pdfsizeopt.py rUNKNOWN.
info: computing statistics for PDF: target.pdf
info: PDF size is 6805076 bytes
info: stat contents_objs = 316509 bytes (5%)
info: stat font_data_objs = 147998 bytes (2%)
info: stat header = 16 bytes (0%)
info: stat image_objs = 5999532 bytes (88%)
info: stat linearized_xref = 17161 bytes (0%)
info: stat other_objs = 317204 bytes (5%)
info: stat separator_data = 27 bytes (0%)
info: stat trailer = 134 bytes (0%)
info: stat wasted_between_objs = 18 bytes (0%)
info: stat xref = 6477 bytes (0%)
info: end of stats