pHash

"The discrete cosine transform (DCT) is an efficient means to compute a hash from frequency spectrum data, and the distance calculation is relatively simple. While it is insufficient to consider image similarity in any semantically meaningful way, it does provide a hash as an ID for an image, and is robust against minor distortions, like small rotations, blurring and compression."

# Compute a 64-bit hash for an image(defdct-img-1(phash/dct-image-hash"image-1.jpg"))(defdct-img-2(phash/dct-image-hash"image-2.png"))# Compare two images via Hamming distance (0 = identical, 64 = completely different)(print(phash/hamming-distancedct-img-1dct-img-2))

"We have developed a new image hash based on the Marr wavelet that computes a perceptual hash based on edge information with particular emphasis on corners. It has been shown that the human visual system makes special use of certain retinal cells to distinguish corner-like stimuli."

(defmh-img-1(phash/mh-image-hash"image.jpg"))(defmh-img-2(phash/dct-image-hash"image-2.png"))(printf"MH hash bytes:\n[1] %s\n[2] %s\n"mh-img-1mh-img-2)(print"Normalized distance: "(phash/hamming-distance2mh-img-1mh-img-2))

"This method tries to take into account geometric features of the image in extracting a hash value. The idea is to generate a feature vector from the variances of 180 lines drawn through the center of the image, and then compact the feature vector with the discrete cosine tranform (DCT). Images can be compared by looking for correlations between the images' hash values. "

(defd(phash/image-digest"image.jpg"))(print"Digest id: "(d:id))# Compare via cross-correlation (1.0 = identical)(defpcc(phash/compare-images"a.jpg""b.jpg"))(print"Cross-correlation: "pcc)

"This hashing method for audio signals extracts a feature vector for every frame of audio from the bark scale frequency spectrum - that is, it only considers those frequencies to which the human auditory system is most sensitive."

(definfo(phash/audio-hash-from-file"recording.wav"))(print"Frames: "(info:count))

"A variable length DCT video hash is included, It consists of the dct image hash applied to a select number of key frames chosen from the original image sequence. The key frames are selected using an adaptive thresholding technique and are based upon a standard framerate, so as not to be fooled by alterations in the frame rate."

(defdct-video-1(phash/dct-video-hash"video.mp4"))(print(lengthdct-video-1)" DCT keyframes extracted")# Compare two videos via minimum Hamming distance across all keyframe pairs(defd(phash/video-distancedct-video-1dct-video-2))(printf"Distance: %d/64"d)

• (dct-image-hash file) → uint64 | nil DCT perceptual hash of an image file.
• (mh-image-hash file &opt alpha lvl) → buffer | nil Marr-Hildreth wavelet hash (72 bytes).
• (mh-image-hash-from-pixels pixels width height channels &opt alpha lvl) → buffer | nil MH hash from raw pixel data.
• (image-digest file &opt sigma gamma n) → {:id string :coeffs buffer :size int} | nil Radial image hash (Radish).
• (compare-images file1 file2 &opt sigma gamma n threshold) → double | nil Cross-correlation between two image digests.
• (cross-correlation digest-x digest-y &opt threshold) → double | nil Cross-correlation between two image digests.

• (text-hash file) → array of {:hash uint64 :index int64} | nil Perceptual text hash points from a file.
• (compare-text-hashes hash1 hash2) → array of {:first int64 :second int64 :length uint32} | nil Find matching text hash sequences between two documents.

• (read-audio file &opt sr channels) → {:samples ptr :length int} | nil Read raw audio samples.
• (read-audio-samples file &opt sr channels) → array of float | nil Read audio samples as a Janet array.
• (audio-hash samples sample-rate) → {:frames buffer :count int} | nil Compute audio hash from raw float samples.
• (audio-hash-from-file file &opt sr channels) → {:frames buffer :count int} | nil Compute audio hash directly from a file.
• (audio-distance-ber hash-a count-a hash-b count-b &opt threshold block-size) → double | nil Bit error rate distance between two audio hashes.

• (dct-video-hash file) → array of uint64 | nil DCT frame hashes from a video file.
• (video-distance hash-a hash-b &opt threshold) → int Minimum Hamming distance (0–64) across all keyframe pairs. Returns 0 when the minimum is ≤ threshold (indicating similar videos). See examples/compare_videos.janet.

• (hamming-distance a b) → int Hamming distance between two 64-bit hashes (0–64).
• (hamming-distance2 hash-a hash-b) → double (0.0–1.0) Normalized Hamming distance between two byte-array hashes.
• (bitcount8 val) → int Number of set bits in a byte (0–8).
• (about) → string pHash version and copyright string.