def model_predict(img_file, view):
"""
:param img_file: image file path
:param view: true if you want show the result
:return:
"""
res_obj = {"score": 0, "position": ""}
prediction = "predict_image.png"
img = cv2.imread(img_file) if view else cv2.imdecode(
numpy.fromstring(img_file, numpy.uint8), 1)
scale = 500 / img.shape[1]
img = cv2.resize(img, (0, 0), fx=scale, fy=scale)
gray, binary = get_binary_image(img)
if get_gray_score(binary):
mg_lbl, regions = selectivesearch.selective_search(
binary, SCALE, SIGMA, MIN_SIZE)
regions = get_proposal(regions, img.shape)
rectangles, score_list = get_prediction(binary, regions)
if len(score_list) > 0:
score = round(max(score_list), 2)
rect = rectangles[score_list.index(max(score_list))]
position = get_pos(rect, scale)
res_obj["score"] = float(score)
res_obj["position"] = position
cv2.drawContours(img, [rect], -1, (255, 145, 30), 2)
if not os.path.exists(PREDICT_PATH):
os.mkdir(PREDICT_PATH)
cv2.imwrite(PREDICT_PATH + prediction, img)
if view:
print(res_obj)
image_view(img)
return res_obj
def create_proposals(image, config):
img_lbl, regions = selectivesearch.selective_search(image,
scale=100,
sigma=0.9,
min_size=10)
candidates = set()
images = []
vertices = []
for r in regions:
# remove reduplicative region
if r['rect'] in candidates:
continue
if r['size'] < config.proposal_min_size:
continue
if (r['rect'][2] * r['rect'][3]) < config.proposal_min_area:
continue
proposal_img, proposal_vertice = clip_pic(image, r['rect'])
if len(proposal_img) == 0:
continue
x, y, w, h = r['rect']
if w == 0 or h == 0:
continue
[a, b, c] = np.shape(proposal_img)
if a == 0 or b == 0 or c == 0:
continue
resized_proposal_img = resize_image(proposal_img,
(config.sz, config.sz))
candidates.add(r['rect'])
img_float = np.asarray(resized_proposal_img, dtype="float32")
images.append(img_float)
vertices.append(r['rect'])
return images, vertices
def searchROI(path):
image = cv2.imread(path)
if image is None:
return
img_lbl, regions = selectivesearch.selective_search(image,
scale=100,
sigma=0.9,
min_size=200)
print(len(regions))
fig, ax = plt.subplots(figsize=(10, 10))
for reg in regions:
x, y, w, h = reg['rect']
rect = cv2.rectangle(image, (x, y), (x + w, y + h), (0, 255, 0), 2)
ax.imshow(rect)
plt.show()
def get_proposals(img):
"""
show generated proposals on input image
:param img: input image
:return:
"""
img = cv2.imread(img)
scale = 500 / img.shape[1]
img = cv2.resize(img, (0, 0), fx=scale, fy=scale)
gray, binary = get_binary_image(img)
mg_lbl, regions = selectivesearch.selective_search(img, min_size=80)
regions = get_proposal(regions, img.shape)
print("proposals:", len(regions))
cv2.drawContours(gray, regions, -1, (255, 145, 30), 2)
image_view(gray)
def read_im(name):
im = Image.open("D:/Statistical Learning/project/image/ours/" + name)
a, b = im.size
if (a > 600):
out = im.resize((800, int((800 * b) / a)), Image.ANTIALIAS)
out.save("D:/Statistical Learning/project/image/ours/" + name)
im = out
im = im.convert('RGB')
img = cv2.imread("D:/Statistical Learning/project/image/ours/" + name)
start_time = time.time()
img_lbl, regions = ss.selective_search(img,
scale=100,
sigma=0.8,
min_size=20)
end_time = time.time()
print('Time for SS: %ssecs' % (end_time - start_time))
return ([im, regions])
def load_train_proposals(datafile,
num_clss,
threshold=0.5,
svm=False,
save=False,
save_path='dataset.pkl'):
train_list = open(datafile, 'r')
labels = []
images = []
for line in train_list:
tmp = line.strip().split(' ')
# tmp0 = image address
# tmp1 = label
# tmp2 = rectangle vertices
img = skimage.io.imread(tmp[0])
img_lbl, regions = selectivesearch.selective_search(img,
scale=500,
sigma=0.9,
min_size=10)
candidates = set()
print(tmp[0])
for r in regions:
# excluding same rectangle (with different segments)
if r['rect'] in candidates:
continue
if r['size'] < 220:
continue
# resize to 224 * 224 for input
proposal_img, proposal_vertice = clip_pic(img, r['rect'])
# Delete Empty array
if len(proposal_img) == 0:
continue
# Ignore things contain 0 or not C contiguous array
x, y, w, h = r['rect']
if w == 0 or h == 0:
continue
# Check if any 0-dimension exist
[a, b, c] = np.shape(proposal_img)
if a == 0 or b == 0 or c == 0:
continue
im = Image.fromarray(proposal_img)
resized_proposal_img = resize_image(im, 224, 224)
candidates.add(r['rect'])
img_float = pil_to_nparray(resized_proposal_img)
images.append(img_float)
# IOU
ref_rect = tmp[2].split(',')
ref_rect_int = [int(i) for i in ref_rect]
iou_val = IOU(ref_rect_int, proposal_vertice)
# labels, let 0 represent default class, which is background
index = int(tmp[1])
if svm == False:
label = np.zeros(num_clss + 1)
if iou_val < threshold:
label[0] = 1
else:
label[index] = 1
labels.append(label)
else:
if iou_val < threshold:
labels.append(0)
else:
labels.append(index)
if save:
pickle.dump((images, labels), open(save_path, 'wb'))
return images, labels
import selectivesearch.selectivesearch as ss
import cv2
img = cv2.imread("./image/1.jpg")
img_lbl, regions = ss.selective_search(img, scale=1000, sigma=0.8, min_size=50)
for r in regions:
x, y, h, w = r['rect']
cv2.rectangle(img, (x, y), (x + h, y + w), (0, 255, 0), 2)
cv2.imshow("img", img)
print("OK")
cv2.waitKey(0)