def templateMatchingDemo(console): root_path = os.path.dirname(os.path.abspath(__file__)) file_path = root_path if console: file_path += "/../../assets/examples/images/square.png" else: file_path += "/../../assets/examples/images/man.jpg" img_color = af.load_image(file_path, True) # Convert the image from RGB to gray-scale img = af.color_space(img_color, af.CSPACE.GRAY, af.CSPACE.RGB) iDims = img.dims() print("Input image dimensions: ", iDims) # Extract a patch from the input image patch_size = 100 tmp_img = img[100:100 + patch_size, 100:100 + patch_size] result = af.match_template(img, tmp_img) # Default disparity metric is # Sum of Absolute differences (SAD) # Currently supported metrics are # AF_SAD, AF_ZSAD, AF_LSAD, AF_SSD, # AF_ZSSD, AF_LSSD disp_img = img / 255.0 disp_tmp = tmp_img / 255.0 disp_res = normalize(result) minval, minloc = af.imin(disp_res) print("Location(linear index) of minimum disparity value = {}".format( minloc)) if not console: marked_res = af.tile(disp_img, 1, 1, 3) marked_res = draw_rectangle(marked_res, minloc%iDims[0], minloc/iDims[0],\ patch_size, patch_size) print( "Note: Based on the disparity metric option provided to matchTemplate function" ) print( "either minimum or maximum disparity location is the starting corner" ) print( "of our best matching patch to template image in the search image") wnd = af.Window(512, 512, "Template Matching Demo") while not wnd.close(): wnd.set_colormap(af.COLORMAP.DEFAULT) wnd.grid(2, 2) wnd[0, 0].image(disp_img, "Search Image") wnd[0, 1].image(disp_tmp, "Template Patch") wnd[1, 0].image(marked_res, "Best Match") wnd.set_colormap(af.COLORMAP.HEAT) wnd[1, 1].image(disp_res, "Disparity Values") wnd.show()
def susan_demo(console): root_path = os.path.dirname(os.path.abspath(__file__)) file_path = root_path if console: file_path += "/../../assets/examples/images/square.png" else: file_path += "/../../assets/examples/images/man.jpg" img_color = af.load_image(file_path, True) img = af.color_space(img_color, af.CSPACE.GRAY, af.CSPACE.RGB) img_color /= 255.0 features = af.susan(img) xs = features.get_xpos().to_list() ys = features.get_ypos().to_list() draw_len = 3 num_features = features.num_features().value for f in range(num_features): print(f) x = xs[f] y = ys[f] # TODO fix coord order to x,y after upstream fix img_color = draw_corners(img_color, y, x, draw_len) print("Features found: {}".format(num_features)) if not console: # Previews color image with green crosshairs wnd = af.Window(512, 512, "SUSAN Feature Detector") while not wnd.close(): wnd.image(img_color) else: print(xs) print(ys)
def load_image(caffe_net, image): tmp = af.load_image(image, is_color=True) af_image = af.resize(tmp, odim0=227, odim1=227) caffe_image = af_image.__array__() py_caffe = '/home/aatish/caffe/python/' t = caffe.io.Transformer({'data': caffe_net.blobs['data'].data.shape}) t.set_mean( 'data', np.load(py_caffe + 'caffe/imagenet/ilsvrc_2012_mean.npy').mean(1).mean(1)) t.set_transpose('data', (2, 0, 1)) t.set_channel_swap('data', (2, 1, 0)) t.set_raw_scale('data', 255.0) caffe_image = t.preprocess('data', af_image.__array__()).reshape(1, 3, 227, 227) caffe_net.blobs['data'].reshape(1, 3, 227, 227) caffe_net.blobs['data'].data[...] = caffe_image af_image = af.np_to_af_array(caffe_image) af_image = af.reorder(af_image, 2, 3, 1, 0) return af_image
# The complete license agreement can be obtained at: # http://arrayfire.com/licenses/BSD-3-Clause ######################################################## import arrayfire as af import sys import os if __name__ == "__main__": if (len(sys.argv) == 1): raise RuntimeError("Expected to the image as the first argument") if not os.path.isfile(sys.argv[1]): raise RuntimeError("File %s not found" % sys.argv[1]) if (len(sys.argv) > 2): af.set_device(int(sys.argv[2])) af.info() hist_win = af.Window(512, 512, "3D Plot example using ArrayFire") img_win = af.Window(480, 640, "Input Image") img = (af.load_image(sys.argv[1])).(af.Dtype.u8) hist = af.histogram(img, 256, 0, 255) while (not hist_win.close()) and (not img_win.close()): hist_win.hist(hist, 0, 255) img_win.image(img)
# The complete license agreement can be obtained at: # http://arrayfire.com/licenses/BSD-3-Clause ######################################################## import arrayfire as af import sys import os if __name__ == "__main__": if (len(sys.argv) == 1): raise RuntimeError("Expected to the image as the first argument") if not os.path.isfile(sys.argv[1]): raise RuntimeError("File %s not found" % sys.argv[1]) if (len(sys.argv) > 2): af.set_device(int(sys.argv[2])) af.info() hist_win = af.Window(512, 512, "3D Plot example using ArrayFire") img_win = af.Window(480, 640, "Input Image") img = af.load_image(sys.argv[1]).as_type(af.Dtype.u8) hist = af.histogram(img, 256, 0, 255) while (not hist_win.close()) and (not img_win.close()): hist_win.hist(hist, 0, 255) img_win.image(img)
def harris_demo(console): root_path = os.path.dirname(os.path.abspath(__file__)) file_path = root_path if console: file_path += "/../../assets/examples/images/square.png" else: file_path += "/../../assets/examples/images/man.jpg" img_color = af.load_image(file_path, True) img = af.color_space(img_color, af.CSPACE.GRAY, af.CSPACE.RGB) img_color /= 255.0 ix, iy = af.gradient(img) ixx = ix * ix ixy = ix * iy iyy = iy * iy # Compute a Gaussian kernel with standard deviation of 1.0 and length of 5 pixels # These values can be changed to use a smaller or larger window gauss_filt = af.gaussian_kernel(5, 5, 1.0, 1.0) # Filter second order derivatives ixx = af.convolve(ixx, gauss_filt) ixy = af.convolve(ixy, gauss_filt) iyy = af.convolve(iyy, gauss_filt) # Calculate trace itr = ixx + iyy # Calculate determinant idet = ixx * iyy - ixy * ixy # Calculate Harris response response = idet - 0.04 * (itr * itr) # Get maximum response for each 3x3 neighborhood mask = af.constant(1, 3, 3) max_resp = af.dilate(response, mask) # Discard responses that are not greater than threshold corners = response > 1e5 corners = corners * response # Discard responses that are not equal to maximum neighborhood response, # scale them to original value corners = (corners == max_resp) * corners # Copy device array to python list on host corners_list = corners.to_list() draw_len = 3 good_corners = 0 for x in range(img_color.dims()[1]): for y in range(img_color.dims()[0]): if corners_list[x][y] > 1e5: img_color = draw_corners(img_color, x, y, draw_len) good_corners += 1 print("Corners found: {}".format(good_corners)) if not console: # Previews color image with green crosshairs wnd = af.Window(512, 512, "Harris Feature Detector") while not wnd.close(): wnd.image(img_color) else: idx = af.where(corners) corners_x = idx / float(corners.dims()[0]) corners_y = idx % float(corners.dims()[0]) print(corners_x) print(corners_y)