def predict_single_image(self,
image,
nms_thresh=0.25,
thickness=2,
multi_colours=True):
if image.shape[0] % 16 or image.shape[1] % 16:
image = cv2.resize(
image, (image.shape[1] // 16 * 16, image.shape[0] // 16 * 16))
img_tensor = torch.from_numpy(
image.transpose((2, 0, 1)).astype(np.float32)).unsqueeze(0)
with torch.no_grad():
out = self.forward(img_tensor.cuda())
out = get_upsample_output(out, self.output_downscale)
pred_dot_map, pred_box_map = get_box_and_dot_maps(
out, nms_thresh, self.BOXES)
index, img_out = get_boxed_img(image,
pred_box_map,
pred_box_map,
pred_dot_map,
self.output_downscale,
self.BOXES,
self.BOX_SIZE_BINS,
thickness=thickness,
multi_colours=multi_colours)
return pred_dot_map, pred_box_map, img_out, index
def predict_single_image(self,
image,
emoji,
nms_thresh=0.25,
thickness=2,
multi_colours=True):
# resize(pad) image for processing if necessary
img_in, w, h = self.handle_image_size(image)
img_tensor = torch.from_numpy(
img_in.transpose((2, 0, 1)).astype(np.float32)).unsqueeze(0)
with torch.no_grad():
out = self.forward(img_tensor)
out = get_upsample_output(out, self.output_downscale)
pred_dot_map, pred_box_map = get_box_and_dot_maps(
out, nms_thresh, self.BOXES)
img_emoji = get_boxed_img(img_in,
emoji,
pred_box_map,
pred_box_map,
pred_dot_map,
self.output_downscale,
self.BOXES,
self.BOX_SIZE_BINS,
thickness=thickness,
multi_colours=multi_colours)
# crop image to original size
img_out = img_emoji[0:h, 0:w]
return pred_dot_map, pred_box_map, img_out
def predict_single_image(self,
image,
nms_thresh=0.25,
thickness=2,
multi_colours=True):
if image.shape[0] % 16 or image.shape[1] % 16:
image = cv2.resize(
image, (image.shape[1] // 16 * 16, image.shape[0] // 16 * 16))
img_tensor = torch.from_numpy(
image.transpose((2, 0, 1)).astype(np.float32)).unsqueeze(0)
with torch.no_grad():
if torch.cuda.is_available():
out = self.forward(img_tensor.cuda())
else:
out = self.forward(img_tensor)
out = get_upsample_output(out, self.output_downscale)
pred_dot_map, pred_box_map = get_box_and_dot_maps(
out, nms_thresh, self.BOXES)
img_out = get_boxed_img(image,
pred_box_map,
pred_box_map,
pred_dot_map,
self.output_downscale,
self.BOXES,
self.BOX_SIZE_BINS,
thickness=thickness,
multi_colours=multi_colours)
# print("type(pred_dot_map):", type(pred_dot_map)) # gt_pred_map
# print("type(pred_box_map):", type(pred_box_map)) # h_map, w_map
# print("len(pred_dot_map):", len(pred_dot_map))
# print("len(pred_box_map):", len(pred_box_map))
# import pdb
# pdb.set_trace()
return pred_dot_map, pred_box_map, img_out