def gen_regions(image, dims, pad):
"""
Generates candidate regions for object detection using selective search
"""
assert(len(dims) == 3)
img = skimage.io.imread(image)
start = time.time()
regions = selective_search(img, ks=[100])
print("Selective search time: %.2f [s]" % (time.time() - start))
resize_t = 0
crops = []
for conf, (x0, y0, x1, y1) in regions:
if x0-pad >= 0:
x0 = x0-pad
if y0-pad >= 0:
y0 = y0-pad
if x1+pad <= dims[0]:
x1 = x1+pad
if y1+pad <= dims[0]:
y1 = y1+pad
region = img[x0:x1, y0:y1, :]
st = time.time()
candidate = resize(region, dims)
resize_t += time.time() - st
crops.append( (conf, candidate, region) )
print("Resize time: %.2f [s]" % resize_t)
return crops
def gen_regions(image, dims, pad, ks):
"""
Generates candidate regions for object detection using selective search.
"""
print "Generating cropped regions..."
assert(len(dims) == 3)
regions = selective_search(image, ks=[ks], feature_masks=[features.SimilarityMask(
size=1,
color=1,
texture=1,
fill=1,
)])
crops = []
for conf, (y0, x0, y1, x1) in regions:
if x0 - pad >= 0:
x0 = x0 - pad
if y0 - pad >= 0:
y0 = y0 - pad
if x1 + pad <= dims[0]:
x1 = x1 + pad
if y1 + pad <= dims[0]:
y1 = y1 + pad
# Images are rows, then columns, then channels.
region = image[y0:y1, x0:x1, :]
candidate = resize(region, dims)
crops.append((conf, candidate, region, (x0, y0, x1, y1)))
print "Generated {} crops".format(len(crops))
return crops
def gen_regions(image, dims, pad):
"""
Generates candidate regions for object detection using selective search
"""
assert (len(dims) == 3)
img = skimage.io.imread(image)
start = time.time()
regions = selective_search(img, ks=[100])
print("Selective search time: %.2f [s]" % (time.time() - start))
resize_t = 0
crops = []
for conf, (x0, y0, x1, y1) in regions:
if x0 - pad >= 0:
x0 = x0 - pad
if y0 - pad >= 0:
y0 = y0 - pad
if x1 + pad <= dims[0]:
x1 = x1 + pad
if y1 + pad <= dims[0]:
y1 = y1 + pad
region = img[x0:x1, y0:y1, :]
st = time.time()
candidate = resize(region, dims)
resize_t += time.time() - st
crops.append((conf, candidate, region))
print("Resize time: %.2f [s]" % resize_t)
return crops
def belltailp_search(self, image):
image_ubyte = img_as_ubyte(image)
candidates = set()
# Actual selective search with some tunable settings, for more
# information look at it's github page
# feature_masks(Texture, Color, Fill, Size)
regions = selective_search(image_ubyte,
color_spaces=['rgb'],
ks=[500, 100],
feature_masks=[(1, 0, 0, 1), (0, 1, 1, 0)])
# Fill in candidates
for region in regions:
size, (y0, x0, y1, x1) = region
candidate = (x0, y0, x1, y1)
candidates.add(candidate)
result = self.render_valid_candidates(image, candidates)
return result
def gen_regions(image, dims, pad, ks):
"""
Generates candidate regions for object detection using selective search.
"""
print "Generating cropped regions..."
assert (len(dims) == 3)
regions = selective_search(image,
ks=[ks],
feature_masks=[
features.SimilarityMask(
size=1,
color=1,
texture=1,
fill=1,
)
])
crops = []
for conf, (y0, x0, y1, x1) in regions:
if x0 - pad >= 0:
x0 = x0 - pad
if y0 - pad >= 0:
y0 = y0 - pad
if x1 + pad <= dims[0]:
x1 = x1 + pad
if y1 + pad <= dims[0]:
y1 = y1 + pad
# Images are rows, then columns, then channels.
region = image[y0:y1, x0:x1, :]
candidate = resize(region, dims)
crops.append((conf, candidate, region, (x0, y0, x1, y1)))
print "Generated {} crops".format(len(crops))
return crops
