This tracker operates in three stages: First, objects are detected in each frame independently using the DPM object detector by Ross Girshick and Pedro Felzenszwalb. Second, all detections with a positive score are associated to detections in the next frame using appearance and the bounding box overlap. We predict objects to the next frame using a Kalman filter and associate them globally via the Hungarian method for bipartite matching. To gap occlusions and missed detections, we also associate tracklets with each other in a first stage. Similarly to the second stage the Hungarian algorithm is employed but this time based on a occlusion sensitive appearance model and regression of the bounding boxes in one tracklet from the bounding boxes in the other tracklet. The algorithm outputs all associated tracklets with a lifetime longer than three frames.
The reported running time is dominated by the object detection stage.
3 GHz, 1 CPU
March 30, 2016 (1 year ago)
March 31, 2016 at 07:34:42 CET
Project page / code:
A. Geiger, M. Lauer, C. Wojek, C. Stiller, R. Urtasun. 3D Traffic Scene Understanding from Movable Platforms. In Pattern Analysis and Machine Intelligence (PAMI), 2014.
|33.7||76.5||1.0||7.2 %||54.2 %||5,804||112,587||2,418||2,252||3 GHz, 1 CPU||Public|
|MOT16-01||22.4||73.7||0.2||23||8.7 %||56.5 %||82||4,814||65||74|
|MOT16-03||39.5||76.5||2.3||148||8.8 %||29.7 %||3,435||58,280||1,509||1,438|
|MOT16-06||38.5||74.7||0.2||221||8.6 %||57.0 %||209||6,692||196||142|
|MOT16-07||27.5||74.8||1.4||54||7.4 %||50.0 %||682||10,938||217||213|
|MOT16-08||25.9||80.7||1.2||63||11.1 %||46.0 %||749||11,477||170||135|
|MOT16-12||36.1||77.8||0.5||86||9.3 %||52.3 %||472||4,754||76||65|
|MOT16-14||13.5||75.9||0.2||164||1.2 %||77.4 %||175||15,632||185||185|