def display(preds, imgs, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
# cv2.imwrite(f'test/img_inferred_d{compound_coef}_this_repo_{i}.jpg', imgs[i])
cv2.imwrite(
f'test_result/d0_epoch71_1031/img_inferred_d{compound_coef}_this_repo_{i}.jpg',
imgs[i])
def display(cur_frame, preds, imgs, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
#cv2.waitKey(0)
if imwrite:
if not os.path.exists(img_path):
os.makedirs(img_path)
cv2.imwrite(
f'{img_path}/img_inferred_d{compound_coef}_this_repo_{cur_frame}.jpg',
imgs[i])
def display(preds, imgs, imshow=True, showtime=0, imwrite=False):
for i, img_name in zip(range(len(imgs)), img_names):
# if len(preds[i]['rois']) == 0: # if model dosen't detect object, not show image
# continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
print(obj)
if imwrite:
img = cv2.cvtColor(imgs[i], cv2.COLOR_BGR2RGB)
cv2.imwrite(f'test/img_inferred_d{compound_coef}_{img_name}', img)
if imshow:
img = cv2.cvtColor(imgs[i], cv2.COLOR_BGR2RGB)
cv2.namedWindow(f'{img_name}', cv2.WINDOW_NORMAL)
cv2.imshow(f'{img_name}', img)
cv2.waitKey(showtime)
cv2.destroyAllWindows()
def display(preds, imgs, imshow=True, imwrite=False, img_id=1):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
imgs[i] = cv2.cvtColor(imgs[i], cv2.COLOR_BGR2RGB)
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
str1 = 'test/' + str(img_id) + '.jpg'
cv2.imwrite(str1, imgs[i])
def display(preds, imgs, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
#----------以下在Colab會crash, 必須改用plt-------------
#cv2.imshow('img', imgs[i])
#cv2.waitKey(0)
#------------------------------------------------------
plt.imshow(imgs[i])
plt.show()
if imwrite:
cv2.imwrite(
f'test/img_inferred_d{compound_coef}_this_repo_{i}.jpg',
imgs[i])
def display(self, preds, imgs, imshow=True, imwrite=False):
scores = []
labels = []
bboxes = []
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = self.system_dict["params"]["obj_list"][preds[i]
['class_ids'][j]]
score = float(preds[i]['scores'][j])
if (score > self.system_dict["params"]["threshold"]):
scores.append(score)
labels.append(obj)
bboxes.append([x1, y1, x2, y2])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=self.system_dict["local"]["color_list"]
[get_index_label(
obj,
self.system_dict["params"]["obj_list"])])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
cv2.imwrite('output.jpg', imgs[i])
return scores, labels, bboxes
def display(preds, imgs, imshow=True, imwrite=False, write_dir=None):
# for i in range(len(imgs)):
for i, (name, img) in enumerate(imgs.items()):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(
img, #imgs[i],
[x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
os.makedirs(write_dir, exist_ok=True)
write_path = os.path.join(write_dir, f'{name}.jpg')
cv2.imwrite(write_path, img)
def display(preds, imgs, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
img = img_path.split('/')[-1].split('.')[0]
# pdb.set_trace()
img_name = f'{img}_img_inferred_d{compound_coef}_this_repo_{i}.jpg'
img_save_path = os.path.join(img_save_folder, img_name)
print(img_save_path)
cv2.imwrite(img_save_path, imgs[i])
def run(img_dir, output_dir, img_size, num_classes, weights, conf_thres,
nms_thres, show):
shutil.rmtree(output_dir, ignore_errors=True)
os.makedirs(output_dir, exist_ok=True)
model = YOLOV3(num_classes, img_size)
state_dict = torch.load(weights, map_location='cpu')
model.load_state_dict(state_dict['model'])
model = model.to(device)
model.eval()
colors = [[random.randint(0, 255) for _ in range(3)] for _ in range(80)]
names = [n for n in os.listdir(img_dir) if osp.splitext(n)[1] in IMG_EXT]
names.sort()
for name in tqdm(names):
img = cv2.imread(osp.join(img_dir, name))
det = inference(model, [img], img_size, conf_thres, nms_thres)[0]
det_txt = []
# Write results
for *xyxy, conf, _, cls in det:
det_txt.append(' '.join(['%g'] * 6) % (*xyxy, cls, conf))
if show: # Add bbox to image
label = '%d %.2f' % (int(cls), conf)
plot_one_box(xyxy, img, label=label, color=colors[int(cls)])
with open(osp.join(output_dir,
osp.splitext(name)[0] + '.txt'), 'w') as f:
f.write('\n'.join(det_txt))
# Stream results
if show:
cv2.imshow('yolo', img)
cv2.waitKey(1)
# Save results (image with detections)
cv2.imwrite(osp.join(output_dir, name), img)
def box(preds, imgs, color_list, obj_list, imshow=False, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2], label=obj,score=score,color=color_list[get_index_label(obj, obj_list)])
return imgs[0]
def draw_result(self, img, det, show=False):
if det is None or len(det) == 0:
if show:
locks.imshow('result', img)
return
for det_pack in det:
xyxy = []
for c in [det_pack.x1, det_pack.y1, det_pack.x2, det_pack.y2]:
xyxy.append(c)
conf = det_pack.class_conf
label = '%s %.2f %.2f' % (det_pack.class_name, conf,
det_pack.object_conf)
plot_one_box(xyxy, img, label=label, color=det_pack.color)
if show:
locks.imshow('result', img)
def display(preds,
imgs,
compound_coef,
obj_list=None,
imshow=True,
imwrite=False,
debug=False):
if obj_list is None:
obj_list = ['person']
color_list = standard_to_bgr(STANDARD_COLORS)
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
if debug:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
os.system("mkdir -p ./assets/predictions")
cv2.imwrite(
f'./assets/predictions/img_inferred_d{compound_coef}_this_repo_{i}.jpg',
imgs[i])
if imwrite:
image_folder = './assets/predictions'
image_files = [
image_folder + '/' + img for img in os.listdir(image_folder)
if img.endswith(".jpg")
]
clip = moviepy.video.io.ImageSequenceClip.ImageSequenceClip(
image_files, fps=1)
clip.write_videofile('./assets/predictions_testset.mp4')
def display(preds, imgs, base_name, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2], label=obj,score=score,color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
cv2.imwrite(f'test003/img_inferred_d{compound_coef}_{base_name}.jpg', imgs[i])
def display(preds, imgs, imwrite=True):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imwrite:
cv2.imwrite(
f'test/img_inferred_d{compound_coef}_this_repo_{i}.jpg',
imgs[i])
def __save_image(self, preds, imgs, imwrite=True):
color_list = standard_to_bgr(STANDARD_COLORS)
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = self.obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(
obj, self.obj_list)])
if imwrite:
cv2.imwrite(
f'test/img_inferred_d{self.compound_coef}_this_repo_{i}.jpg',
imgs[i])
def display(preds, imgs, imshow=True, imwrite=False):
for i, img_name in zip(range(len(imgs)), img_names):
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2], label=obj, score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
img = cv2.cvtColor(imgs[i], cv2.COLOR_BGR2RGB)
cv2.imshow('video', img)
cv2.waitKey(1)
if imwrite:
cv2.imwrite(f'test/img_inferred_d{compound_coef}_{img_name}', imgs[i])
def display(preds, imgs, img_path, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2], label=obj,score=score,color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
write_path = img_path.split(".")[0] + "_predicted.jpg";
cv2.imwrite(write_path, imgs[i]);
def draw(self, frames_dict, preds_dict, judgements_dict, fps):
vis_imgs_dict = {}
for name in preds_dict.keys():
frame = frames_dict[name]
pred = preds_dict[name]
judgement = judgements_dict[name]
img = self.draw_static_contents(frame, name)
img = self.draw_fps(img, fps)
label = 'person'
if pred is not None:
for x1, y1, x2, y2 in pred:
box_area = (x2 - x1) * (y2 - y1)
# 过滤掉过大和过小的识别框
if min_object_bbox_area_dict[
name] <= box_area <= max_object_bbox_area_dict[
name]:
plot_one_box((x1, y1, x2, y2),
img,
label=label,
color=(225, 225, 0))
if judgement:
img = cv2.putText(img,
text='Kick your head!!!',
org=(30, 25),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=1.2,
color=(0, 0, 255),
thickness=2)
else:
img = cv2.putText(img,
text='Safe working',
org=(30, 25),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=1.2,
color=(0, 255, 0),
thickness=2)
vis_imgs_dict[name] = img
return vis_imgs_dict
def display(preds, imgs, outdir, files, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
out_name = outdir + files[i]
cv2.imwrite(out_name, imgs[i])
def display(out_1, out_2, imgs, imshow=True, showtime=0, imwrite=False):
# if len(preds[i]['rois']) == 0: # if model dosen't detect object, not show image
# continue
for img, out_1 in zip(imgs, out_1):
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
for i in range(len(out_1['rois'])):
ox1, oy1, ox2, oy2 = out_1['rois'][i].astype(np.int)
obj_1 = obj_list_1[out_1['class_ids'][i]]
score = float(out_1['scores'][i])
color = color_list[get_index_label(obj_1, obj_list_1)]
plot_one_box(img, [ox1, oy1, ox2, oy2],
label=obj_1,
score=score,
color=color)
print(obj_1)
print(f'obj {i}:', ox1, oy1, ox2, oy2)
for j in range(len(out_2[i]['rois'])):
dx1, dy1, dx2, dy2 = out_2[i]['rois'][j].astype(np.int)
obj_2 = obj_list_2[out_2[i]['class_ids'][j]].split('_')[-1]
score = float(out_2[i]['scores'][j])
plot_one_box(img, [dx1 + ox1, dy1 + oy1, dx2 + ox1, dy2 + oy1],
label=obj_2,
score=score,
color=color)
print(obj_2)
print('damage :', dx1, dy1, dx2, dy2)
print('change :', dx1 + ox1, dy1 + oy1, dx2 + ox1, dy2 + oy1)
if imshow:
# print(f'{img_name}')
cv2.namedWindow('__', cv2.WINDOW_NORMAL)
cv2.resizeWindow('__', 1500, 1000)
cv2.imshow('__', img)
# cv2.waitKey(0)
key = cv2.waitKey(showtime)
if key == ord('p'):
cv2.waitKey(-1)
def display(names, preds, imgs, imshow=True, imwrite=False, save_bbox=True):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
f = open('test/' + names[i] + '.txt', "w")
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if save_bbox and obj == 'cell phone':
f.write(
str(67) + ' ' + str(score) + ' ' + str(x1) + ' ' +
str(y1) + ' ' + str(x2) + ' ' + str(y2) + '\n')
if imshow:
cv2.imshow(names[i], imgs[i])
cv2.waitKey(0)
if imwrite:
cv2.imwrite('test/' + names[i] + '.jpg', imgs[i])
def display(preds, imgs, imshow=True, imwrite=True):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
plot_one_box(imgs[i], [x1, y1, x2, y2], label=obj, score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
imgs[i] = cv2.cvtColor(imgs[i], cv2.COLOR_BGR2RGB)
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
cv2.imwrite(f'test/d{compound_coef}_{haha}', imgs[i])
print('nmsl')
def analyse_rgb():
t = time.time()
img0 = get_bgr()
img_org = img0[:, :, ::-1]
img, _, _, _ = letterbox(img0, new_shape=image_size)
img = img[:, :, ::-1].transpose(2, 0, 1) # BGR to RGB
img = np.ascontiguousarray(img, dtype=np.float32)
img /= 255.0
img = torch.from_numpy(img).unsqueeze(0).to('cpu')
pred, _ = model(img)
det = non_max_suppression(pred, 0.6, 0.5)[0]
if det is not None and len(det) > 0:
detected_classes = []
print('+ Rescaling model')
det[:, :4] = scale_coords(img.shape[2:], det[:, :4], img0.shape).round()
for *coordinates, conf, cls_conf, cls in det:
if classes[int(cls)] in RISKY_CLASSES:
label = '%s %.2f' % (classes[int(cls)], conf)
plot_one_box(coordinates, img0, label=label, color=colors[int(cls)])
print(f"+ Detected {classes[int(cls)]}")
detected_classes.append({classes[int(cls)]: {'x': coordinates[0], 'y': coordinates[1]}})
n = []
for counter in detected_classes:
width = img0.shape[1]
x, y = counter[list(counter.keys())[0]].values()
phi = (x / width * 2 - 1) * (CAMERA_FOV / 2)
n.append(f"{list(counter.keys())[0]};{phi};-1|")
s = str(''.join(str(x) for x in n)[:-1])
return {"raw": get_rgb(), "done": img0, "objects": s}
return {'raw': img_org, 'done': img_org, 'objects': ''}
def display(preds, imgs, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
imgs[i] = imgs[i].copy()
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
list_image_id.append(img_path.split('/')[-1])
list_class.append(obj)
list_score.append(score)
list_xmin.append(x1)
list_ymin.append(y1)
list_xmax.append(x2)
list_ymax.append(y2)
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
def display(preds, imgs, imshow=True, imwrite=False):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
print(x1, y1, x2, y2, obj, score)
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
predicated_path = 'demo_jpg/{0}_infer.jpg'.format(
config.dataset_name)
cv2.imwrite(predicated_path, imgs[i])
print("write predicated result in:{0}".format(predicated_path))
def display(preds, imgs, imshow=True, imwrite=False, start=0):
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
if obj in obj_interest:
plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(obj, obj_list)])
if imshow:
cv2.imshow('img', imgs[i])
cv2.waitKey(0)
if imwrite:
cv2.imwrite(
img_root + 'Result' +
f'/img_inferred_d{compound_coef}_this_repo_{i+start}.jpg',
imgs[i])
def display(preds, imgs, imshow=True, imwrite=False):
x = []
y = []
result_1 = []
result_2 = []
for i in range(len(imgs)):
if len(preds[i]['rois']) == 0:
continue
for j in range(len(preds[i]['rois'])):
x1, y1, x2, y2 = preds[i]['rois'][j].astype(np.int)
obj = obj_list[preds[i]['class_ids'][j]]
score = float(preds[i]['scores'][j])
x, y = plot_one_box(imgs[i], [x1, y1, x2, y2],
label=obj,
score=score,
color=color_list[get_index_label(
obj, obj_list)])
result_1.append(x)
result_2.append(y)
return result_1, result_2
def __find_intersections_v1(self):
for flag_idx in self.class_det["Flag"]:
# 1. Collect total number of intersection with detected Person object
detected_intersection = []
for person_idx in self.class_det["Person"]:
flag_xyxy = get_det_xyxy(self.det[flag_idx])
person_xyxy = get_det_xyxy(self.det[person_idx])
if not self.opt.maximize_latency:
if self.plot_old_person_bbox: # Plot in bbox
plot_one_box(person_xyxy,
self.img_mbbox,
label="Person-%s" % str(person_idx),
color=self.rgb["MMBox"])
person_xyxy = self.__enlarge_bbox(
person_xyxy) # enlarge bbox size
# Intersection occurs here
if self.__is_intersect(flag_xyxy, person_xyxy):
detected_intersection.append(person_idx)
# Testing only: Try plotting bounding boxes
if not self.opt.maximize_latency:
plot_one_box(person_xyxy,
self.img_enlarge,
label="EnPer-%s" % str(person_idx),
color=self.rgb["EnlargedPerson"])
plot_one_box(flag_xyxy,
self.img_enlarge,
label="Flag-%s" % str(flag_idx),
color=self.rgb["Person"])
self.__verify_intersection(flag_idx, detected_intersection)
# save MB-Box illustration
# print(" >>>>> self.save_path = ", self.save_path)
# cv2.imwrite(self.save_path+, self.img_mbbox)
cv2.imwrite(
self.save_path.replace('.png', '') + "-mbbox.png", self.img_mbbox)
cv2.imwrite(
self.save_path.replace('.png', '') + "-enlarge.png",
self.img_enlarge)
def detect(model, dataset, args):
use_cuda = not args.cpu
threshold = args.threshold
iou_threshold = args.iou_threshold
input_sizes = [512, 640, 768, 896, 1024, 1280, 1280, 1536, 1536]
input_size = input_sizes[args.compound_coef]
img_dir = os.path.join(dataset, dataset, 'images')
bbox_dir = os.path.join(dataset, dataset, 'annotations', 'bboxes')
vis_dir = os.path.join(dataset, 'det_vis')
prepare_dirs(bbox_dir, vis_dir)
img_paths = [os.path.join(img_dir, f) for f in os.listdir(img_dir)]
for img_path in tqdm(img_paths):
ori_imgs, framed_imgs, framed_metas = preprocess(img_path,
max_size=input_size)
ori_img = ori_imgs[0]
img_id = os.path.basename(img_path).split('.')[0]
json_byhand = os.path.join(dataset, 'annotation_byhand',
img_id + '.json')
if os.path.exists(json_byhand):
with open(json_byhand) as f:
annotation_byhand = json.load(f)
points = annotation_byhand['shapes'][0]['points']
max_box = points[0] + points[1]
else:
if args.update: # only process annotations by hand
continue
if use_cuda:
x = torch.stack(
[torch.from_numpy(fi).cuda() for fi in framed_imgs], 0)
else:
x = torch.stack([torch.from_numpy(ft) for fi in framed_imgs],
0)
x = x.to(torch.float32).permute(0, 3, 1, 2)
with torch.no_grad():
features, regression, classification, anchors = model(x)
regressBoxes = BBoxTransform()
clipBoxes = ClipBoxes()
preds = postprocess(x, anchors, regression, classification,
regressBoxes, clipBoxes, threshold,
iou_threshold)
pred = invert_affine(framed_metas, preds)[0]
max_area, max_box = 0, [0, 0, ori_img.shape[1], ori_img.shape[0]]
for det, class_id in zip(pred['rois'], pred['class_ids']):
if not class_id == 0:
continue
x1, y1, x2, y2 = det.astype(np.int)
w, h = x2 - x1, y2 - y1
area = w * h
if area > max_area:
max_area = area
max_box = [x1, y1, x2, y2]
plot_one_box(ori_img, max_box, color=[255, 0, 255], line_thickness=2)
if args.vis:
cv2.imwrite(os.path.join(vis_dir, img_id + '.jpg'), ori_img)
bbox_file = os.path.join(bbox_dir, img_id + '.txt')
with open(bbox_file, 'w') as f:
bbox_info = ' '.join(map(str, max_box))
f.write(bbox_info)
from detector import YoloV3Predictor
import sys
from PIL import Image
import numpy as np
import json
import random
from utils.utils import plot_one_box
if __name__ == '__main__':
model = YoloV3Predictor(0, 416, half_precision=False)
if len(sys.argv) == 1:
img = np.array(Image.open('data/samples/bus.jpg'))
else:
img = np.array(Image.open(sys.argv[1]))
res = model.predict(img, conf_thres=0.5, agnostic_iou=False)
with open('cfg/labels.json', 'r') as f:
label_dict = json.load(f)
colors = [[random.randint(0, 255) for _ in range(3)]
for _ in range(len(label_dict))]
for *xyxy, conf, cls in res[0]:
label = '%s %.2f' % (list(label_dict.keys())[int(cls)], conf)
plot_one_box(xyxy, img, label=label, color=colors[int(cls)])
Image.fromarray(img).save('output/result.jpg')