Detects colors in images 8 x faster than Numpy / Uses Cython - returns a dict
Project description
Detects colors in images 8 x faster than Numpy / Uses Cython - returns a dict
Tested against Windows 10 / Python 3.11 / Anaconda
pip install locate-pixelcolor-cython-multi-auto-compile
Very fast RGB color search in pictures.
I wrote a couple of variations of this function. All of them can be used in Python.
Cython, detects colors in images 2-3 x faster than Numpy # pre-compiled for Python 3.10 https://github.com/hansalemaos/locate_pixelcolor_cythonsingle
Cython, but with multiple processors (5-10x faster than Numpy) # pre-compiled for Python 3.10 https://github.com/hansalemaos/locate_pixelcolor_cythonmulti
Cupy, using the GPU (up to 8x faster than Numpy) https://github.com/hansalemaos/locate_pixelcolor_cupy
C - shared library (10x faster than Numpy) https://github.com/hansalemaos/locate_pixelcolor_c
C++ - parallel_for - shared library (up to 10x faster than Numpy) https://github.com/hansalemaos/locate_pixelcolor_cpp_parallelfor
C++ - pragma omp - shared library (20x faster than Numpy) https://github.com/hansalemaos/locate_pixelcolor_cpppragma
Numba - compiled - ahead of time (2-3x faster than numpy) https://github.com/hansalemaos/locate_pixelcolor_numba
Numba Cuda - compiled - ahead of time (10x faster than numpy) https://github.com/hansalemaos/locate_pixelcolor_numbacuda
import numpy as np
import cv2
from locate_pixelcolor_cython_multi_auto_compile import search_colors
# 4525 x 6623 x 3 picture https://www.pexels.com/pt-br/foto/foto-da-raposa-sentada-no-chao-2295744/
picx = r"C:\Users\hansc\Downloads\pexels-alex-andrews-2295744.jpg"
pic = cv2.imread(picx)
colors0 = np.array([[255, 255, 255]],dtype=np.uint8)
resus0 = search_colors(pic=pic, colors=colors0)
colors1=np.array([(66, 71, 69),(62, 67, 65),(144, 155, 153),(52, 57, 55),(127, 138, 136),(53, 58, 56),(51, 56, 54),(32, 27, 18),(24, 17, 8),],dtype=np.uint8)
resus1 = search_colors(pic=pic, colors=colors1)
# b,g,r = pic[...,0],pic[...,1],pic[...,2]
# %timeit resus0=search_colors(pic,colors0)
# %timeit np.where(((b==255)&(g==255)&(r==255)))
# %timeit resus1=search_colors(pic, colors1)
# %timeit np.where(((b==66)&(g==71)&(r==69))|((b==62)&(g==67)&(r==65))|((b==144)&(g==155)&(r==153))|((b==52)&(g==57)&(r==55))|((b==127)&(g==138)&(r==136))|((b==53)&(g==58)&(r==56))|((b==51)&(g==56)&(r==54))|((b==32)&(g==27)&(r==18))|((b==24)&(g==17)&(r==8)))
# 22.9 ms ± 63.9 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)
# 161 ms ± 1.96 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
# 131 ms ± 400 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)
# 939 ms ± 6.95 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
# resus1
# Out[5]:
# defaultdict(list,
# {(66, 71, 69): array([[ 0, 4522],
# [ 3, 4522],
# [ 3, 4523],
# ...,
# [5895, 852],
# [6043, 2871],
# [6063, 2897]], dtype=int32),
# (51,
# 56,
# 54): array([[ 2, 4522],
# [ 15, 4523],
# [ 48, 4035],
# ...,
# [5898, 1961],
# [5898, 1962],
# [5903, 1975]], dtype=int32),
# (127,
# 138,
# 136): array([[ 0, 38],
# [ 2, 0],
# [ 3, 19],
# [ 5, 13],
# [ 6, 10],
# [ 8, 9],
# [ 8, 42],
# [ 9, 34],
# [ 9, 46],
# [ 11, 1],
# [ 11, 8],
# [ 11, 9],
# [ 11, 25],
# [ 13, 0],
# [ 14, 36],
# [ 15, 21],
# [ 15, 22],
# [ 15, 33],
# [ 15, 40],
# [ 49, 56],
# [ 562, 3300],
# [ 615, 1746],
# [ 616, 1738],
# [ 662, 15],
# [ 663, 11],
# [ 894, 3218],
# [ 897, 3249],
# [ 907, 2166],
# [ 937, 1673],
# [1241, 3230],
# [1245, 1691],
# [1246, 1693],
# [1318, 1705],
# [1319, 1651],
# [1366, 1590],
# [1371, 1536],
# [1372, 1541],
# [1379, 1586],
# [1379, 1623],
# [1382, 1589],
# [1383, 1594],
# [1384, 1623],
# [1386, 1564],
# [1391, 1584],
# [1394, 1569],
# [1394, 1571],
# [1394, 1573],
# [1398, 1591],
# [1400, 1569],
# [1404, 1551],
# [1404, 1571],
# [1405, 1555],
# [1406, 1566],
# [1412, 1585],
# [1414, 1550],
# [1414, 1560],
# [1414, 1567],
# [1415, 1538],
# [1416, 1568],
# [1420, 1562],
# [1420, 1573],
# [1421, 1575],
# [1423, 1554],
# [1438, 1518],
# [1479, 3069],
# [2295, 2404],
# [2310, 2346],
# [2312, 2371],
# [2323, 2353],
# [2323, 2714],
# [2391, 2672],
# [2642, 3440],
# [2851, 2706],
# [2863, 2625],
# [2921, 2383],
# [2926, 2633],
# [2932, 2651],
# [2951, 2685],
# [2953, 2688],
# [2972, 2464],
# [3148, 2606],
# [3295, 2509],
# [3311, 2617],
# [3314, 2622],
# [3320, 2623],
# [3325, 2368],
# [3331, 2613],
# [3332, 2614],
# [3337, 2310],
# [3343, 2615],
# [3372, 2300],
# [3375, 2595],
# [3382, 2597],
# [4208, 2576],
# [4227, 2577],
# [4268, 289],
# [4347, 433],
# [4564, 811],
# [4686, 884],
# [4752, 860],
# [4995, 856],
# [5013, 2974],
# [5050, 460],
# [5093, 3457],
# [5094, 2925],
# [5130, 2849],
# [5131, 2850],
# [5136, 666],
# [5169, 3548],
# [5180, 3337],
# [5247, 3101],
# [5256, 2947],
# [5257, 3232],
# [5265, 3250],
# [5275, 2935],
# [5298, 2866],
# [5314, 3369],
# [5321, 3859],
# [5363, 2916],
# [5364, 3182],
# [5394, 2728],
# [5399, 3344],
# [5434, 2579],
# [5448, 3321],
# [5452, 3678],
# [5476, 3328],
# [5509, 2973]], dtype=int32),
# (144,
# 155,
# 153): array([[ 1, 1],
# [ 1, 40],
# [ 2, 21],
# [ 6, 5],
# [ 6, 22],
# [ 6, 25],
# [ 6, 69],
# [ 8, 22],
# [ 9, 1],
# [ 9, 38],
# [ 9, 66],
# [ 9, 75],
# [ 10, 38],
# [ 11, 17],
# [ 11, 22],
# [ 12, 17],
# [ 13, 37],
# [ 16, 4],
# [ 16, 91],
# [ 17, 81],
# [ 17, 94],
# [ 19, 14],
# [ 20, 0],
# [ 21, 2],
# [ 48, 60],
# [ 63, 57],
# [ 877, 3035],
# [ 969, 2294],
# [1018, 2260],
# [1258, 3185],
# [1258, 3189],
# [1262, 3218],
# [1262, 3219],
# [1263, 1673],
# [1277, 3212],
# [1303, 1717],
# [1312, 1660],
# [1337, 1641],
# [1340, 1642],
# [1360, 1583],
# [1363, 3163],
# [1366, 1573],
# [1369, 1596],
# [1370, 1591],
# [1372, 1573],
# [1381, 1620],
# [1382, 1594],
# [1393, 1599],
# [1395, 1592],
# [1397, 1596],
# [1400, 1563],
# [1400, 1592],
# [1402, 1561],
# [1403, 1544],
# [1403, 1554],
# [1404, 1560],
# [1405, 1603],
# [1407, 1586],
# [1409, 1563],
# [1409, 1590],
# [1409, 1596],
# [1410, 1597],
# [1411, 1556],
# [1411, 1599],
# [1420, 1577],
# [1797, 2667],
# [2151, 2318],
# [2293, 2347],
# [2297, 2351],
# [2297, 2368],
# [2301, 2335],
# [2301, 2381],
# [2306, 2371],
# [2306, 2376],
# [2309, 2416],
# [2309, 2424],
# [2317, 2363],
# [2318, 2371],
# [2319, 2412],
# [2321, 2425],
# [2326, 2366],
# [2329, 2401],
# [2383, 2601],
# [2687, 2636],
# [2710, 2731],
# [2718, 2689],
# [2810, 2597],
# [2968, 2457],
# [2970, 2442],
# [3004, 2684],
# [3010, 2413],
# [3070, 2515],
# [3075, 2524],
# [3087, 2669],
# [3209, 2529],
# [3300, 2524],
# [3330, 2613],
# [3331, 2514],
# [3347, 2381],
# [3465, 2568],
# [3888, 2430],
# [3942, 1891],
# [3942, 1893],
# [3950, 1890],
# [4227, 2576],
# [4321, 98],
# [4420, 497],
# [4420, 500],
# [4421, 500],
# [4496, 619],
# [4530, 814],
# [4721, 859],
# [4723, 858],
# [4770, 952],
# [4822, 3101],
# [4864, 3195],
# [4923, 155],
# [4941, 548],
# [4942, 549],
# [4978, 2675],
# [4993, 882],
# [5112, 3145],
# [5141, 625],
# [5149, 736],
# [5185, 643],
# [5217, 3420],
# [5225, 3391],
# [5226, 3386],
# [5227, 3673],
# [5237, 3260],
# [5299, 4207],
# [5308, 3709],
# [5320, 3857],
# [5328, 3750],
# [5359, 2744],
# [5365, 2847],
# [5372, 2779],
# [5388, 2603],
# [5389, 3079],
# [5392, 2871],
# [5399, 2879],
# [5416, 2895],
# [5424, 3085],
# [5428, 3107],
# [5435, 2580],
# [5457, 3375],
# [5461, 3093],
# [5466, 3252],
# [5466, 3254],
# [5472, 3356],
# [5478, 671],
# [5483, 3150],
# [5485, 3670],
# [5487, 3672],
# [5510, 3433],
# [5513, 3107],
# [5537, 3010],
# [5537, 3012],
# [5579, 2977],
# [5755, 1061]], dtype=int32),
# (32,
# 27,
# 18): array([[ 3, 3240],
# [ 3, 3241],
# [ 12, 3349],
# ...,
# [6622, 2844],
# [6622, 2854],
# [6622, 2865]], dtype=int32),
# (24,
# 17,
# 8): array([[ 669, 1723],
# [ 670, 1722],
# [ 781, 3120],
# ...,
# [6622, 4522],
# [6622, 4523],
# [6622, 4524]], dtype=int32),
# (53,
# 58,
# 56): array([[ 2, 4523],
# [ 2, 4524],
# [ 4, 4522],
# ...,
# [5906, 1969],
# [5908, 1246],
# [6057, 2862]], dtype=int32),
# (52,
# 57,
# 55): array([[ 1, 4524],
# [ 49, 4035],
# [ 52, 4030],
# ...,
# [6040, 2868],
# [6049, 2868],
# [6058, 2863]], dtype=int32),
# (62,
# 67,
# 65): array([[ 0, 4523],
# [ 50, 4037],
# [ 52, 4033],
# ...,
# [5903, 1973],
# [5904, 1952],
# [6041, 2868]], dtype=int32)})
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file locate_pixelcolor_cython_multi_auto_compile-0.10.tar.gz.
File metadata
- Download URL: locate_pixelcolor_cython_multi_auto_compile-0.10.tar.gz
- Upload date:
- Size: 27.6 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.11.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | d216120c78d1a184a448c422658d58ec70bda6003b90339bcc36c1cf4060f538 |
|
| MD5 | 2b61b4623401717c8f09233759251b67 |
|
| BLAKE2b-256 | 96a7bb3848f94de00e6edd1f14db0b5e9b57538a2ee165cc94dfe15776430a04 |
File details
Details for the file locate_pixelcolor_cython_multi_auto_compile-0.10-py3-none-any.whl.
File metadata
- Download URL: locate_pixelcolor_cython_multi_auto_compile-0.10-py3-none-any.whl
- Upload date:
- Size: 27.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.11.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 423adca4eb605815f9e477e6ab21f9509f9f0b19015778475e5bc81f70fd2405 |
|
| MD5 | b45015c08426ab793440d5fce722cca7 |
|
| BLAKE2b-256 | 7040aa720fa51a7393528a3de531b789ac5305564e94a70b0a3bcf979ffe45f6 |
