******************************************************************************* Instruction for Treadmill Dataset A (2010/03/24) Yagi laboratory, Department of Intelligent Media The Institute of Scientific and Industrial Research, Osaka University http://www.am.sanken.osaka-u.ac.jp/index.html ******************************************************************************* ------------------------------------------------------------------------------- 1. Contents ------------------------------------------------------------------------------- The Treadmill Dataset A is composed of 34 subjects with speed variation from 2 km/h to 7 km/h at 1 km/h interval. Each subject have two sequences of "gallery" and "probe". The directory structure is as follows. TreadmillDatasetA -(5-digits subject ID) -(gallery/probe)_(speed)km -(8-digits image ID).png e.g. ./TreadmillDatasetA/00000/gallery_2km/00000001.png The number of frames for each sequence is as follows. 2km/h: 420 3km/h: 360 4km/h: 360 5km/h: 420 6km/h: 360 7km/h: 240 ------------------------------------------------------------------------------- 1. Data specification ------------------------------------------------------------------------------- The data specification for training and testing is as follows. Basically, the training set is picked up only from gallery, while the test set is picked up both from gallery and probe. ------------ + Subset 1 ------------ Training subject IDs 001 003 008 055 094 095 096 098 099 100 101 104 106 111 Test subject IDs 112 116 117 118 124 126 128 132 133 134 136 137 140 143 145 146 147 148 149 150 ------------ + Subset 2 ------------ Training subject IDs 001 003 008 055 094 095 096 098 099 Test subject IDs 112 116 117 118 124 126 128 132 133 134 136 137 140 143 145 146 147 148 149 150 --------------------------------------------------------------- + Subset for comparison with related works (related to Fig. 7) --------------------------------------------------------------- Training subject IDs 003 055 096 100 101 112 116 137 146 Test subject IDs 001 008 094 095 098 099 104 106 111 117 118 124 126 128 132 133 134 136 140 143 145 147 148 149 150