| commit | author | age | ||
| 9efc18 | 1 | import cv2 |
| SP | 2 | import numpy as np |
| 3 | from skimage import morphology,img_as_ubyte | |
| 02e0f7 | 4 | from sklearn import svm |
| SP | 5 | from sklearn.externals import joblib |
| 6 | ||
| 7 | ||
| 9efc18 | 8 | |
| SP | 9 | """ |
| 10 | (1) The text is an array of chars (in row-major order) where | |
| 11 | * each char can be one of the following: | |
| 12 | * 'x': hit | |
| 13 | * 'o': miss | |
| 14 | * ' ': don't-care | |
| 15 | * (2) When the origin falls on a hit or miss, use an upper case | |
| 16 | * char (e.g., 'X' or 'O') to indicate it. When the origin | |
| 17 | * falls on a don't-care, indicate this with a 'C'. | |
| 18 | * The string must have exactly one origin specified. | |
| 19 | * (3) The advantage of this method is that the text can be input | |
| 20 | * in a format that shows the 2D layout of the Sel; e.g., | |
| 21 | ||
| 22 | ||
| 23 | :::: AND :::: | |
| 24 | ||
| 25 | ||
| 26 | (10) The sequence string is formatted as follows: | |
| 27 | * ~ An arbitrary number of operations, each separated | |
| 28 | * by a '+' character. White space is ignored. | |
| 29 | * ~ Each operation begins with a case-independent character | |
| 30 | * specifying the operation: | |
| 31 | * d or D (dilation) | |
| 32 | * e or E (erosion) | |
| 33 | * o or O (opening) | |
| 34 | * c or C (closing) | |
| 35 | * r or R (rank binary reduction) | |
| 36 | * x or X (replicative binary expansion) | |
| 37 | * b or B (add a border of 0 pixels of this size) | |
| 38 | * ~ The args to the morphological operations are bricks of hits, | |
| 39 | * and are formatted as a.b, where a and b are horizontal and | |
| 40 | * vertical dimensions, rsp. | |
| 41 | * ~ The args to the reduction are a sequence of up to 4 integers, | |
| 42 | * each from 1 to 4. | |
| 43 | * ~ The arg to the expansion is a power of two, in the set | |
| 44 | * {2, 4, 8, 16}. | |
| 45 | * (11) An example valid sequence is: | |
| 46 | * "b32 + o1.3 + C3.1 + r23 + e2.2 + D3.2 + X4" | |
| 47 | * In this example, the following operation sequence is carried out: | |
| 48 | * * b32: Add a 32 pixel border around the input image | |
| 49 | * * o1.3: Opening with vert sel of length 3 (e.g., 1 x 3) | |
| 50 | * * C3.1: Closing with horiz sel of length 3 (e.g., 3 x 1) | |
| 51 | * * r23: Two successive 2x2 reductions with rank 2 in the first | |
| 52 | * and rank 3 in the second. The result is a 4x reduced pix. | |
| 53 | * * e2.2: Erosion with a 2x2 sel (origin will be at x,y: 0,0) | |
| 54 | * * d3.2: Dilation with a 3x2 sel (origin will be at x,y: 1,0) | |
| 55 | * * X4: 4x replicative expansion, back to original resolution | |
| 56 | ||
| 57 | """ | |
| 58 | ||
| 59 | ||
| 60 | def kernel(x, y): | |
| 61 | return np.ones((x, y), np.uint8) | |
| 62 | ||
| 63 | ||
| 02e0f7 | 64 | def getSID(image, classifier): |
| 9efc18 | 65 | image=255-image |
| SP | 66 | image=img_as_ubyte(image>100) |
| 67 | cv2.imwrite("enSID0.png", image) | |
| 68 | # Remove noise | |
| 69 | image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(2,2), iterations=1) | |
| 70 | # Closing. Connect non connected parts | |
| 02e0f7 | 71 | image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel(5, 3), iterations=4) |
| 9efc18 | 72 | # Again noise removal after closing |
| 02e0f7 | 73 | |
| 9efc18 | 74 | image = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel(8,8), iterations=1) |
| SP | 75 | # Skeletonization |
| 76 | image = img_as_ubyte(morphology.thin(image>128)) | |
| 77 | cv2.imwrite("enSID1.png",image) | |
| 78 | # Stub removal (might not be necessary if thinning instead of skeletonize is used above | |
| 79 | # Making lines stronger | |
| 80 | image = cv2.morphologyEx(image, cv2.MORPH_DILATE, kernel(5, 5), iterations=1) | |
| 81 | ||
| 82 | image = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel(10, 10)) | |
| 83 | # Thining again | |
| 84 | image = img_as_ubyte(morphology.skeletonize(image>0.5)) | |
| 85 | image = cv2.morphologyEx(image, cv2.MORPH_DILATE, kernel(10, 10)) | |
| 02e0f7 | 86 | |
| SP | 87 | im2,ctrs, hier = cv2.findContours(image.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) |
| 88 | sorted_ctrs = sorted(ctrs, key=lambda ctr: cv2.boundingRect(ctr)[0]) | |
| 89 | ||
| 90 | #classifier = joblib.load('filename.joblib') | |
| 91 | ||
| 92 | sid_no="" | |
| 93 | for i, ctr in enumerate(sorted_ctrs): | |
| 94 | # Get bounding box | |
| 95 | x, y, w, h = cv2.boundingRect(ctr) | |
| 96 | # Getting ROI | |
| 97 | if(w<h/2): | |
| 98 | sid_no=sid_no+"1" | |
| 99 | continue | |
| 100 | roi = image[y:y+h, x:x+w] | |
| 101 | roi = img_as_ubyte(roi < 128) | |
| 102 | roi = cv2.resize(roi,(32,32)) | |
| 103 | ||
| 104 | #cv2.rectangle(image,(x,y),( x + w, y + h ),(0,255,0),2) | |
| 105 | cv2.imwrite('sid_no_{}.png'.format(i), roi) | |
| 106 | sid_no=sid_no+str(classifier.predict(roi.reshape(1,-1)/255.0)[0]) | |
| 107 | print(sid_no) | |
| 9efc18 | 108 | return image |
