Lucid Data Dreaming for Multiple Object Tracking
Authors:
- Anna KhorevaAKAnna Khoreva
- Rodrigo BenensonRBRodrigo Benenson
- Eddy IlgEIEddy Ilg
- Thomas BroxTBThomas Brox
- Bernt SchieleBSBernt Schiele
Preprint
Abstract
Convolutional networks reach top quality in pixel-level object tracking but require a large amount of training data (1k~10k) to deliver such results. We propose a new training strategy which achieves state-of-the-art results across three evaluation datasets while using 20x~100x less annotated data than competing methods. Our approach is suitable for both single and multiple object tracking. Instead of using large training sets hoping to generalize across domains, we generate in-domain training data using the provided annotation on the first frame of each video to synthesize ("lucid dream") plausible future video frames. In-domain per-video training data allows us to train high quality appearance- and motion-based models, as well as tune the post-processing stage. This approach allows to reach competitive results even when training from only a single annotated frame, without ImageNet pre-training. Our results indicate that using a larger training set is not automatically better, and that for the tracking task a smaller training set that is closer to the target domain is more effective. This changes the mindset regarding how many training samples and general "objectness" knowledge are required for the object tracking task.Keywords
All Related Versions
This publication has 50 references indexed in Scilit:
- Wider or Deeper: Revisiting the ResNet Model for Visual RecognitionPattern Recognition, 2019
- Back to Basics: Unsupervised Learning of Optical Flow via Brightness Constancy and Motion Smoothness2016
- Fully-Convolutional Siamese Networks for Object Tracking2016
- Playing for Data: Ground Truth from Computer Games2016
- Normalized Cut Meets MRF2016
- Learning to Track at 100 FPS with Deep Regression Networks2016
- The Pascal Visual Object Classes Challenge: A RetrospectiveInternational Journal of Computer Vision, 2014
- TouchCut: Fast image and video segmentation using single-touch interactionComputer Vision and Image Understanding, 2014
- Supervoxel-Consistent Foreground Propagation in Video2014
- Exploiting the Circulant Structure of Tracking-by-Detection with Kernels2012