distance.pl

#!/usr/bin/env perl

use strict;
use warnings;
use feature qw(say);
use experimental qw(signatures);

my $THRESHOLD = 8;

sub distance($a, $b) {
    my $count = 0;
    my $xor = ($a+0) ^ ($b+0);
    while($xor > 0) {
        if($xor & 1) {
            $count++;
        }
        $xor >>= 1;
    }

    return $count;
}

my @entries;

while(<>) {
    chomp;
    my ( $filename, $hash ) = split /\t/, $_;

    next unless -e $filename;

    push @entries, [ $filename, $hash ];
}

for my $i (0..$#entries) {
    for my $j ($i+1..$#entries) {
        my $d = distance($entries[$i][1], $entries[$j][1]);
        if($d <= $THRESHOLD) {
            say join("\t", $entries[$i][0], $entries[$j][0], $d);
        }
    }
}

phash.cc

#include <opencv2/opencv.hpp>
#include <opencv2/highgui.hpp>

#include <cstdint>
#include <cstdio>
#include <string>
#include <unistd.h>

using namespace cv;
using namespace std;

// http://www.hackerfactor.com/blog/?/archives/432-Looks-Like-It.html

static
float median(const Mat &img)
{
    Mat copy = img.clone().reshape(0, 64);
    cv::sort(copy, copy, cv::SORT_EVERY_COLUMN | cv::SORT_ASCENDING);
    return (copy.at<float>(31) + copy.at<float>(32)) / 2.0;
}

static uint64_t
phash(const Mat &img)
{
    Mat eight_by_eight;
    Mat grayscale;
    Mat float_grayscale;
    Mat discrete_cosine;

    resize(img, eight_by_eight, Size(32, 32));
    cvtColor(eight_by_eight, grayscale, cv::COLOR_BGR2GRAY);
    grayscale.convertTo(float_grayscale, CV_32FC1);
    dct(float_grayscale, discrete_cosine);
    Mat top8(discrete_cosine, Rect(1, 1, 8, 8)); // original cuts off the first row and column
    //top8.at<float>(0, 0) = 0;
    double avg = median(top8);
    Mat avg_mat(8, 8, CV_32FC1, Scalar(avg, 0, 0, 0));
    avg_mat = (top8 > avg_mat);

    uint64_t mask = 1UL << 63;
    uint64_t result = 0;
    for(int row = 0; row < 8; row++) {
        for(int col = 0; col < 8; col++) {
            if(avg_mat.at<uchar>(row, col) != 0) {
                result |= mask;
            }
            mask >>= 1;
        }
    }
    return result;
}

int
main(int argc, char **argv)
{
    for(int i = 1; i < argc; i++) {
        if(access(argv[i], R_OK)) {
            continue;
        }

        try {
            Mat img = imread(argv[i]);
            uint64_t hash = phash(img);
            printf("%s\t%llu\n", argv[i], hash);
        } catch(...) {
            fprintf(stderr, "exception for %s\n", argv[i]);
        }
    }
    return 0;
}