analysis_results.py

import _init_paths
import matplotlib.pyplot as plt
plt.rcParams['figure.figsize'] = [8, 8]
import argparse
from lib.test.analysis.plot_results import plot_results, print_results, print_per_sequence_results
from lib.test.evaluation import get_dataset, trackerlist

def parse_args():
    """
    args for training.
    """
    parser = argparse.ArgumentParser(description='Parse args for training')
    # for evaluation
    parser.add_argument('--name', type=str, help='model name')

    args = parser.parse_args()

    return args

if __name__ == "__main__":
    trackers = []
    dataset_name = 'GOT10K'

    args = parse_args()
    trackers.extend(trackerlist(name='unicorn_sot', parameter_name=args.name, dataset_name=dataset_name,
                                run_ids=None, display_name='Unicorn'))

    dataset = get_dataset('got10k_val')
    # plot_results(trackers, dataset, 'OTB2015', merge_results=True, plot_types=('success', 'norm_prec'),
    #              skip_missing_seq=False, force_evaluation=True, plot_bin_gap=0.05)
    print_results(trackers, dataset, dataset_name, merge_results=True, plot_types=('success', 'prec', 'norm_prec'))
    # print_per_sequence_results(trackers, dataset, "Unicorn")