def run_bench(delete_after=False, args_ext=None):
if args_ext == None:
args = parser.parse_args()
else:
args = args_ext
cfg = load_cfg(args)
cfg['THOR']['viz'] = args.viz
cfg['THOR']['verbose'] = args.verbose
# setup tracker and dataset
if args.tracker == 'SiamFC':
tracker = SiamFC_Tracker(cfg)
elif args.tracker == 'SiamRPN':
tracker = SiamRPN_Tracker(cfg)
elif args.tracker == 'SiamMask':
tracker = SiamMask_Tracker(cfg)
else:
raise ValueError(f"Tracker {args.tracker} does not exist.")
dataset = load_dataset(args.dataset)
# optionally filter for a specific videos
if args.spec_video:
dataset = {args.spec_video: dataset[args.spec_video]}
if args.dataset == "VOT2018":
test_bench, eval_bench = test_vot, eval_vot
elif args.dataset == "OTB2015":
test_bench, eval_bench = test_otb, eval_otb
else:
raise NotImplementedError(
f"Procedure for {args.dataset} does not exist.")
# testing
total_lost = 0
speed_list = []
for v_id, video in enumerate(dataset.keys(), start=1):
tracker.temp_mem.do_full_init = True
speed = test_bench(v_id, tracker, dataset[video], args)
speed_list.append(speed)
# evaluation
mean_fps = np.mean(np.array(speed_list))
bench_res = eval_bench(args.save_path, delete_after)
bench_res['mean_fps'] = mean_fps
print(bench_res)
return bench_res
def load_cfg(args):
json_path = f"configs/{args.tracker}/VOT2018_"
if args.vanilla:
json_path += "vanilla.json"
else:
json_path += f"THOR_{args.lb_type}.json"
cfg = json.load(open(json_path))
return cfg
if __name__ == '__main__':
args = parser.parse_args()
cfg = load_cfg(args)
cfg['THOR']['viz'] = args.viz
cfg['THOR']['verbose'] = args.verbose
print("[INFO] Initializing the tracker.")
if args.tracker == 'SiamFC':
tracker = SiamFC_Tracker(cfg)
elif args.tracker == 'SiamRPN':
tracker = SiamRPN_Tracker(cfg)
elif args.tracker == 'SiamMask':
tracker = SiamMask_Tracker(cfg)
# elif args.tracker == 'SiamRPN_PP':
# tracker = SiamRPN_PP_Tracker(cfg)
else:
raise ValueError(f"Tracker {args.tracker} does not exist.")
print("[INFO] Starting video stream.")
show_webcam(tracker, mirror=True, viz=args.viz)
def run_bench(delete_after=False):
args = parser.parse_args()
cfg = load_cfg(args)
cfg['THOR']['viz'] = args.viz
cfg['THOR']['verbose'] = args.verbose
# setup tracker and dataset
if args.tracker == 'SiamFC':
tracker = SiamFC_Tracker(cfg)
elif args.tracker == 'SiamRPN':
tracker = SiamRPN_Tracker(cfg)
elif args.tracker == 'SiamMask':
tracker = SiamMask_Tracker(cfg)
else:
raise ValueError(f"Tracker {args.tracker} does not exist.")
dataset = load_dataset(args.dataset)
# optionally filter for a specific videos
if args.spec_video:
# pdb.set_trace()
dataset = {args.spec_video: dataset[args.spec_video]}
if args.dataset == "VOT2018":
test_bench, eval_bench = test_vot, eval_vot
elif args.dataset == "OTB2015":
test_bench, eval_bench = test_otb, eval_otb
elif args.dataset == "GOT10k":
test_bench = test_got
elif args.dataset == "GOT10k_train_val":
test_bench = test_gottrainval
elif args.dataset == "LaSOT":
test_bench, eval_bench = test_lasot, eval_lasot
elif args.dataset == "UAV20L":
test_bench, eval_bench = test_uav20l, eval_uav20l
elif args.dataset == "UAV123":
test_bench, eval_bench = test_uav123, eval_uav123
elif args.dataset == "OXUVA":
test_bench = test_oxuva
elif args.dataset == "TC128":
test_bench, eval_bench = test_tc128, eval_tc128
else:
raise NotImplementedError(
f"Procedure for {args.dataset} does not exist.")
# testing
total_lost = 0
speed_list = []
if args.dataset == "OTB2015":
print("==>> No processing for the json file ... ")
else:
dataset = ast.literal_eval(dataset)
# pdb.set_trace()
for v_id, video in enumerate(dataset.keys(), start=1):
tracker.temp_mem.do_full_init = True
speed = test_bench(v_id, tracker, dataset[video], args)
speed_list.append(speed)
if args.dataset == "GOT10k":
print("$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$")
print("==>> Please evaluate online for GOT10k dataset ... ")
print("$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$")
elif args.dataset == "OxUvA":
print("$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$")
print("==>> Please evaluate online for OxUvA dataset ... ")
print("$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$")
else:
# evaluation
# pdb.set_trace()
bench_res = eval_bench(args.save_path, delete_after)
print(bench_res)
mean_fps = np.mean(np.array(speed_list))
bench_res['mean_fps'] = mean_fps
print(bench_res)
return bench_res