KAIST
A widely used public dataset in the RGB-T object detection field. It is relatively large in scale and contains synchronized visible-thermal image pairs captured by multiple sensors, covering scenarios such as daytime and nighttime.
Timeline (last 90 days)
Under controlled RGB-T corruption simulation, PG-FM maintains performance on clean datasets while improving mAP50 under corruption conditions by up to 20.8 points on the FLIR dataset and up to 19.9 points on the KAIST dataset.
The KAIST dataset is larger and contains visible-thermal image pairs synchronously collected by multiple sensors, covering various scenes such as daytime, nighttime, campus, and street, with labeled categories including pedestrians, cyclists, and vehicles.
The DARPA Robotics Challenge (DRC) was launched by the Defense Advanced Research Projects Agency in 2012; in the 2015 finals, KAIST of South Korea won by completing all 8 tasks with DRC-HUBO in the fastest time.
DRC最终赛于2015年在加州波莫纳举行，韩国KAIST团队的DRC-HUBO以最快时间完成所有任务夺冠，获得200万美元奖金
波士顿动力的Atlas和韩国KAIST的DRC-HUBO在DARPA机器人挑战赛中一战成名
All Facts (5)
Under controlled RGB-T corruption simulation, PG-FM maintains performance on clean datasets while improving mAP50 under corruption conditions by up to 20.8 points on the FLIR dataset and up to 19.9 points on the KAIST dataset.
50%UnverifiedThe KAIST dataset is larger and contains visible-thermal image pairs synchronously collected by multiple sensors, covering various scenes such as daytime, nighttime, campus, and street, with labeled categories including pedestrians, cyclists, and vehicles.
50%UnverifiedDRC最终赛于2015年在加州波莫纳举行，韩国KAIST团队的DRC-HUBO以最快时间完成所有任务夺冠，获得200万美元奖金
50%Unverified波士顿动力的Atlas和韩国KAIST的DRC-HUBO在DARPA机器人挑战赛中一战成名
50%UnverifiedThe DARPA Robotics Challenge (DRC) was launched by the Defense Advanced Research Projects Agency in 2012; in the 2015 finals, KAIST of South Korea won by completing all 8 tasks with DRC-HUBO in the fastest time.
50%