About ROAD
The ROAD dataset is a pioneer in action classification on the road. It consists of labeled videos of agents in road scenes, perfect for training a neural-based action classifier. Below shows a depiction of ROAD’s 3 sub-tasks:
As of now, the Watonomous Action Classification team is concerned with tasks 1 and 2, object detection/tracking and action detection/tracking. Semantic location classification is also an interesting task in and of itself, but its implementation into the car is vaguer and arguably not needed. Keeping this in mind, the technical goal of Action Classification is to produce a neural network capable of doing action detection/tracking (ROAD task 2) to a high degree of precision and recall for each of the action classes. Note that achieving task 2 may also require the completion of task 1 as action classification is a high-level task that requires a preconceived understanding of the world (in this case, an object detector/tracker).
For file: road_trainval_v1.0.json, fill as we need
'all_input_labels':
['AV', 'Car', 'SmalVeh', 'MedVeh', 'LarVeh', 'Bus', 'Mobike', 'Cyc', 'Ped', 'TL', 'OthTL', 'EmVeh', 'Red', 'Amber', 'Green', 'black', 'MovAway', 'MovTow', 'Mov', 'Rev', 'Brake', 'Stop', 'IncatLft', 'IncatRht', 'HazLit', 'TurLft', 'TurRht', 'MovRht', 'MovLft', 'Ovtak', 'Wait2X', 'XingFmLft', 'XingFmRht', 'Xing', 'PushObj', 'OutgoBusLane', 'IncomBusLane', 'OutgoCycLane', 'IncomCycLane', 'VehLane', 'OutgoLane', 'IncomLane', 'LftPav', 'RhtPav', 'Pav', 'Jun', 'xing', 'BusStop', 'parking', 'LftParking', 'rightParking']
'all_av_action_labels':
['AV-Stop', 'AV-Mov', 'AV-TurRht', 'AV-TurLft', 'AV-MovRht', 'AV-MovLft', 'AV-Ovtak', 'AV-black']
'av_action_labels':
['AV-Stop', 'AV-Mov', 'AV-TurRht', 'AV-TurLft', 'AV-MovRht', 'AV-MovLft', 'AV-Ovtak']
'agent_labels':
['Ped', 'Car', 'Cyc', 'Mobike', 'MedVeh', 'LarVeh', 'Bus', 'EmVeh', 'TL', 'OthTL']
'action_labels':
['Red', 'Amber', 'Green', 'MovAway', 'MovTow', 'Mov', 'Brake', 'Stop', 'IncatLft', 'IncatRht', 'HazLit', 'TurLft', 'TurRht', 'Ovtak', 'Wait2X', 'XingFmLft', 'XingFmRht', 'Xing', 'PushObj']
'duplex_labels':
['Bus-MovAway', 'Bus-MovTow', 'Bus-Stop', 'Bus-XingFmLft', 'Car-Brake', 'Car-IncatLft', 'Car-IncatRht', 'Car-MovAway', 'Car-MovTow', 'Car-Stop', 'Car-TurLft', 'Car-TurRht', 'Car-XingFmLft', 'Car-XingFmRht', 'Cyc-MovAway', 'Cyc-MovTow', 'Cyc-Stop', 'Cyc-TurLft', 'Cyc-XingFmLft', 'Cyc-XingFmRht', 'LarVeh-Stop', 'MedVeh-IncatLft', 'MedVeh-MovTow', 'MedVeh-Stop', 'MedVeh-TurRht', 'OthTL-Green', 'OthTL-Red', 'Ped-Mov', 'Ped-MovAway', 'Ped-MovTow', 'Ped-PushObj', 'Ped-Stop', 'Ped-Wait2X', 'Ped-Xing', 'Ped-XingFmLft', 'Ped-XingFmRht', 'TL-Amber', 'TL-Green', 'TL-Red']
'triplet_labels'
'loc_labels'
'old_loc_labels'
'db':
| '2014-06-25-16-45-34_stereo_centre_02':
| '2014-06-26-09-53-12_stereo_centre_02':
| '2014-07-14-14-49-50_stereo_centre_01':
| | 'split_ids':
| | | ['all', 'train_1', 'train_2', 'val_3']
| | 'agent_tubes':
| | 'action_tubes':
| | 'loc_tubes':
| | 'duplex_tubes':
| | 'triplet_tubes':
| | 'av_action_tubes':
| | 'frame_labels':
| | 'frames':
| | | '1':
| | | '2':
| | | '3':
| | | | 'annotated': 1
| | | | 'rgb_image_id': 3
| | | | 'width': 1280
| | | | 'height': 960
| | | | 'av_action_ids': [1]
| | | | 'annos':
| | | | | b05:
| | | | | | 'box': [0.4950002338420697, 0.4074358974358974, 0.5347014207947216, 0.47469975611363785]
| | | | | | 'agent_ids': [6]
| | | | | | 'loc_ids': [0]
| | | | | | 'action_ids': [8, 11]
| | | | | | 'duplex_ids': [91, 94]
| | | | | | 'triplet_ids': [1092, 1128]
| | | | | | 'tube_uid': '8258d065'
| | | | | b06:
| | | | | ...
| | | | 'input_image_id': 3 # to my understanding, same as rgb_image_id
| | | ...
| | 'numf':
| ...
'label_types'
'all_duplex_labels'
'all_triplet_labels'
'all_agent_labels'
'all_loc_labels'
'all_action_labels'
'duplex_childs'
'triplet_childs'