Module corpus 

Corpus steganography: zero-modification cover selection via ANN search.

Given a payload to embed and a corpus of images, this module finds the image whose natural LSB bit pattern most closely matches the payload — minimising or eliminating modifications needed.

Functions

extract_lsb_pattern

Extract the LSB bit pattern from raw pixel bytes.

filter_by_model

Filter corpus entries by model ID and resolution.

hamming_distance

Compute the Hamming distance between two byte slices of equal length.

is_close_match

Determine if a Hamming distance counts as a “close enough” match.

payload_to_bit_pattern

Expand a payload into a bit pattern of the same format as extract_lsb_pattern output (one bit per sample, packed into bytes).

score_match

Score a corpus entry’s precomputed bit pattern against a payload pattern.