acia.segm.processor.offline#

Offline image processors to perform segmentation

class acia.segm.processor.offline.OfflineModel[source]#

Model that runs on the local computer

__init__(config_file, parameter_file, half=False, device='cuda', tiling=None)[source]#: config_file: model configuration file parameter_file: model checkpoint file half: enables half-precision (16-bit) execution. A bit faster. device: chooses the device to execute (e.g. ‘cpu’ or ‘cuda’ or ‘cuda:0’)

load_model(device=None, cfg_options=None, half=False)[source]#

Load model from definitions

device: device type, e.g. ‘cpu’ or ‘cuda’ cfg_options: overwrite configuration options e.g. {‘test_cfg.rpn.nms_thr’: 0.7}

predict(source)[source]#

Predicts the overlay for an image sequence

source: image sequence source tiling: whether to enable tiling

class acia.segm.processor.offline.PoseModel[source]#

Model that runs on the local computer

__init__(model_name='bact_omni', omni=True, use_gpu=torch.cuda.is_available, diameter=None, flow_threshold=None)[source]#: config_file: model configuration file parameter_file: model checkpoint file half: enables half-precision (16-bit) execution. A bit faster. device: chooses the device to execute (e.g. ‘cpu’ or ‘cuda’ or ‘cuda:0’)

predict(source)[source]#

Predicts the overlay for an image sequence

source: image sequence source tiling: whether to enable tiling