Adapting Vehicle Detector to Target Domain by Adversarial Prediction Alignment
Y Koga, H Miyazaki, R Shibasaki - 2021 IEEE International …, 2021 - ieeexplore.ieee.org
Y Koga, H Miyazaki, R Shibasaki
2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021•ieeexplore.ieee.orgWhile recent advancement of domain adaptation techniques is significant, most of methods
only align a feature extractor and do not adapt a classifier to target domain, which would be
a cause of performance degradation. We propose a novel domain adaptation technique for
object detection that aligns prediction output space. In addition to feature alignment, we
aligned predictions of locations and class confidences of our vehicle detector for satellite
images by adversarial training. The proposed method significantly improved AP score by …
only align a feature extractor and do not adapt a classifier to target domain, which would be
a cause of performance degradation. We propose a novel domain adaptation technique for
object detection that aligns prediction output space. In addition to feature alignment, we
aligned predictions of locations and class confidences of our vehicle detector for satellite
images by adversarial training. The proposed method significantly improved AP score by …
While recent advancement of domain adaptation techniques is significant, most of methods only align a feature extractor and do not adapt a classifier to target domain, which would be a cause of performance degradation. We propose a novel domain adaptation technique for object detection that aligns prediction output space. In addition to feature alignment, we aligned predictions of locations and class confidences of our vehicle detector for satellite images by adversarial training. The proposed method significantly improved AP score by over 5%, which shows effectivity of our method for object detection tasks in satellite images. Our code is available at https://github.com/monotaro3/vd_pred_align.
ieeexplore.ieee.org