MMCV-HandGestureRecognition/featureBasedSift.py at main · Czyzo66/MMCV-HandGestureRecognition · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
import math

import cv2
import numpy as np
import os
from dataclasses import dataclass

@dataclass
class ClassficationResult:
    classNames: []
    goodMatchList: []
    success: bool
    resultString: str
    matches: []
    kp1: []
    kp2: []
    img1: object
    accuracy: str


class FeatureBasedClassifier:

    #Initialize detector
    def __init__(self):
        self.sift = cv2.SIFT_create()
        self.images = []
        self.classNames = []
        self.kp1 = []
        self.kp2 = []
        self.matches = []
        self.allMatchCount = 0
        self.goodMatchList = []

    def loadImageClasses(self):
        #Loading image classes
        print(os.listdir(os.curdir))
        path = 'FeatureBased/Classes'
        # images = []
        # classNames = []

        classList = os.listdir(path)
        classList.remove(".DS_Store")
        print('Classes detected: ',len(classList))

        for imageClass in classList:
            image = cv2.imread(f'{path}/{imageClass}',0)
            self.images.append(image)
            self.classNames.append(os.path.splitext(imageClass)[0])
        print(self.classNames)

    #create descriptiors for the classes
    def getDescriptors(self,imageList):
        descriptorList=[]
        for image in imageList:
            kp,des = self.sift.detectAndCompute(image,None)
            descriptorList.append(des)
            self.kp1.append(kp)
        return descriptorList

    #classify the image to the class
    def getId(self,image, descriptorList):
        #find descriptor of the image to be classified
        kp,des = self.sift.detectAndCompute(image,None)
        self.kp2.append(kp)
        bruteForceMatcher = cv2.BFMatcher()
        matchList = []
        imageClassifiedToIndex = -1
        minNumberOfMatches = 2
        try:
            for descriptor in descriptorList:
                bruteForceMatches = bruteForceMatcher.knnMatch(descriptor,des, k=2)
                bruteForceMatches2 = bruteForceMatcher.match(descriptor,des)
                goodMatches = []
                for i,j in bruteForceMatches:
                    if i.distance < 0.75 * j.distance:
                        goodMatches.append(i)
                matchList.append(len(goodMatches))
                self.allMatchCount+= len(goodMatches)
                self.matches.append(goodMatches)
        except:
            pass
        if len(matchList) != 0:
            if max(matchList) > minNumberOfMatches:
                imageClassifiedToIndex = matchList.index(max(matchList))
        print(matchList)
        self.goodMatchList.append(matchList)

        return imageClassifiedToIndex

    def getAccuracy(self, id):
        mean = self.allMatchCount / len(self.goodMatchList[0])
        sum = 0
        for matchCount in self.goodMatchList[0]:
            sum += pow((matchCount - mean),2)

        deviation = math.sqrt(sum/self.allMatchCount)

        accuracy = self.goodMatchList[0][id] - mean
        accurayDividedByDeviation = accuracy / deviation

        return accurayDividedByDeviation

    def cleanup(self):
        self.kp1 = []
        self.kp2 = []
        self.matches = []
        self.allMatchCount = 0
        self.goodMatchList = []

    def classify(self, image):
        descriptorList = self.getDescriptors(self.images)
        # print(len(descriptorList))
        id = self.getId(image, descriptorList)

        if id != -1:
            accuracy= str(round(self.getAccuracy(id),2))
            result = ClassficationResult(success=True,
                                         resultString="Result: " + self.classNames[id],
                                         matches=self.matches[id],
                                         kp1=self.kp1[id],
                                         kp2=self.kp2[0],
                                         img1=self.images[id],
                                         accuracy=str(accuracy),
                                         classNames=self.classNames,
                                         goodMatchList=self.goodMatchList
                                         )
            self.cleanup()
            return result
        else:
            result = ClassficationResult(success=False,resultString="Result: Unknown")
            self.cleanup()
            return result