Generally the number of bits is the base 2 logarithm of the number of possible symbols (e.g. photos, words, characters) raised to the power of the number of random symbols in use.
So 128 symbols that can either be a 1 or a 0 is log2(2^128) = 128 bits.
And 10 words from a 7776-word dictionary is log2(7776^10), or approximately equal to 128.
I think it'll be lot harder for you to memorize photos than words.