mmdetection/demo/image_demo.py

# Copyright (c) OpenMMLab. All rights reserved.
"""Image Demo.

This script adopts a new infenence class, currently supports image path,
np.array and folder input formats, and will support video and webcam
in the future.

Example:
    Save visualizations and predictions results::

        python demo/image_demo.py demo/demo.jpg rtmdet-s

        python demo/image_demo.py demo/demo.jpg \
        configs/rtmdet/rtmdet_s_8xb32-300e_coco.py \
        --weights rtmdet_s_8xb32-300e_coco_20220905_161602-387a891e.pth

        python demo/image_demo.py demo/demo.jpg \
        glip_atss_swin-t_a_fpn_dyhead_pretrain_obj365 --texts bench

        python demo/image_demo.py demo/demo.jpg \
        glip_atss_swin-t_a_fpn_dyhead_pretrain_obj365 --texts 'bench . car .'

        python demo/image_demo.py demo/demo.jpg \
        glip_atss_swin-t_a_fpn_dyhead_pretrain_obj365
        --texts 'bench . car .' -c

        python demo/image_demo.py demo/demo.jpg \
        glip_atss_swin-t_a_fpn_dyhead_pretrain_obj365 \
        --texts 'There are a lot of cars here.'

    Visualize prediction results::

        python demo/image_demo.py demo/demo.jpg rtmdet-ins-s --show

        python demo/image_demo.py demo/demo.jpg rtmdet-ins_s_8xb32-300e_coco \
        --show
"""

from argparse import ArgumentParser

from mmengine.logging import print_log

from mmdet.apis import DetInferencer


def parse_args():
    parser = ArgumentParser()
    parser.add_argument(
        'inputs', type=str, help='Input image file or folder path.')
    parser.add_argument(
        'model',
        type=str,
        help='Config or checkpoint .pth file or the model name '
        'and alias defined in metafile. The model configuration '
        'file will try to read from .pth if the parameter is '
        'a .pth weights file.')
    parser.add_argument('--weights', default=None, help='Checkpoint file')
    parser.add_argument(
        '--out-dir',
        type=str,
        default='outputs',
        help='Output directory of images or prediction results.')
    parser.add_argument('--texts', help='text prompt')
    parser.add_argument(
        '--device', default='cuda:0', help='Device used for inference')
    parser.add_argument(
        '--pred-score-thr',
        type=float,
        default=0.3,
        help='bbox score threshold')
    parser.add_argument(
        '--batch-size', type=int, default=1, help='Inference batch size.')
    parser.add_argument(
        '--show',
        action='store_true',
        help='Display the image in a popup window.')
    parser.add_argument(
        '--no-save-vis',
        action='store_true',
        help='Do not save detection vis results')
    parser.add_argument(
        '--no-save-pred',
        action='store_true',
        help='Do not save detection json results')
    parser.add_argument(
        '--print-result',
        action='store_true',
        help='Whether to print the results.')
    parser.add_argument(
        '--palette',
        default='none',
        choices=['coco', 'voc', 'citys', 'random', 'none'],
        help='Color palette used for visualization')
    # only for GLIP
    parser.add_argument(
        '--custom-entities',
        '-c',
        action='store_true',
        help='Whether to customize entity names? '
        'If so, the input text should be '
        '"cls_name1 . cls_name2 . cls_name3 ." format')

    call_args = vars(parser.parse_args())

    if call_args['no_save_vis'] and call_args['no_save_pred']:
        call_args['out_dir'] = ''

    if call_args['model'].endswith('.pth'):
        print_log('The model is a weight file, automatically '
                  'assign the model to --weights')
        call_args['weights'] = call_args['model']
        call_args['model'] = None

    init_kws = ['model', 'weights', 'device', 'palette']
    init_args = {}
    for init_kw in init_kws:
        init_args[init_kw] = call_args.pop(init_kw)

    return init_args, call_args


def main():
    init_args, call_args = parse_args()
    # TODO: Video and Webcam are currently not supported and
    #  may consume too much memory if your input folder has a lot of images.
    #  We will be optimized later.
    inferencer = DetInferencer(**init_args)
    inferencer(**call_args)

    if call_args['out_dir'] != '' and not (call_args['no_save_vis']
                                           and call_args['no_save_pred']):
        print_log(f'results have been saved at {call_args["out_dir"]}')


if __name__ == '__main__':
    main()