图片相似度

1. 余弦相似度

from PIL import Image
from numpy import average, dot, linalg

# 对图片进行统一化处理
def get_thum(img_path, size=(64, 64), greyscale=False):
    img = Image.open(img_path)
    # 对图像大小重置, Image.ANTIALIAS为高质量的
    img = img.resize(size, Image.ANTIALIAS)
    if greyscale:
        # 将图片转换为L模式，其为灰度图，其每个像素用8个bit表示
        img = img.convert('L')
    return img

# 计算图片的余弦距离
def image_cosin_similarity(img1_path, img2_path):
    img1 = get_thum(img1_path)
    img2 = get_thum(img2_path)
    imgs = [img1, img2]
    vectors = []
    norms = []
    for img in imgs:
        vector = []
        for pixel_tuple in img.getdata():
            vector.append(average(pixel_tuple))
        vectors.append(vector)
        # linalg=linear（线性）+algebra（代数），norm则表示范数
        norms.append(linalg.norm(vector, 2))
    # dot返回的是点积，对二维数组（矩阵）进行计算
    res = dot(vectors[0] / norms[0], vectors[1] / norms[1])
    return res

2. 其他计算

2.1 汉明距离

def hamming_distance(str1, str2):
    if len(str1) != len(str2):
        return
    count = 0
    for i in range(len(str1)):
        if str1[i] != str2[i]:
            count += 1
    return count

2.2 计算md5值

import hashlib

def md5_score(img_path='22.png'):
    with open(img_path, "rb") as f:
        pmd5 = hashlib.md5(f.read())
    return pmd5.hexdigest()

2.3 哈希算法

2.3.1 均值哈希算法

import cv2

def ahash(img_path):
    image = cv2.imread(img_path='22.png')
    # 将图片缩放为8*8的
    image = cv2.resize(image, (8,8), interpolation=cv2.INTER_CUBIC)
    # 将图片转化为灰度图
    gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
    # s为像素和初始灰度值，hash_str为哈希值初始值
    s = 0
    ahash_str = ''
    # 遍历像素累加和
    for i in range(8):
        for j in range(8):
            s = s + gray[i, j]
    # 计算像素平均值
    avg = s / 64
    # 灰度大于平均值为1相反为0，得到图片的平均哈希值，此时得到的hash值为64位的01字符串
    ahash_str = ''
    for i in range(8):
        for j in range(8):
            if gray[i, j] > avg:
                ahash_str = ahash_str + '1'
            else:
                ahash_str = ahash_str + '0'
    result = ''
    for i in range(0, 64, 4):
        result += ''.join('%x' % int(ahash_str[i: i + 4], 2))
    # print("ahash值：",result)
    return result

2.3.2 差异值哈希算法

import cv2

def dhash(img_path):
    image = cv2.imread(img_path)
    # 将图片转化为8*8
    image = cv2.resize(image, (9, 8), interpolation=cv2.INTER_CUBIC)
    # 将图片转化为灰度图
    gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
    dhash_str = ''
    for i in range(8):
        for j in range(8):
            if gray[i, j] > gray[i, j + 1]:
                dhash_str = dhash_str + '1'
            else:
                dhash_str = dhash_str + '0'
    result = ''
    for i in range(0, 64, 4):
        result += ''.join('%x' % int(dhash_str[i: i + 4], 2))
    # print("dhash值",result)
    return result

2.3.3 计算hash值

import cv2

def phash(img_path):
    image = cv2.imread(img_path)
    # 将图片转化为8*8
    image = cv2.resize(image, (8, 8), interpolation=cv2.INTER_CUBIC)
    # 将图片转化为灰度图
    gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
    # 计算图片的平均灰度值
    avg = sum([sum(gray[i]) for i in range(8)]) / 64
    # 计算哈希值,与平均值比较生成01字符串
    str = ''
    for i in range(8):
        str += ''.join(map(lambda i: '0' if i < avg else '1', gray[i]))
    # 计算hash值, 将64位的hash值，每4位合成以为，转化为16 位的hash值
    result = ''
    for i in range(0, 64, 4):
        result += ''.join('%x' % int(str[i: i + 4], 2))
    # print(result)
    return result

3. 深度神经网络提取图片向量，计算相似度

检索出与指定图片相似度最高的n张图片