Lucid Data Dreaming for Multiple Object Tracking

03/28/2017
by   Anna Khoreva, et al.
0

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.

READ FULL TEXT

page 1

page 3

page 5

page 6

page 12

page 13

page 16

page 17

research
02/10/2019

MOTS: Multi-Object Tracking and Segmentation

This paper extends the popular task of multi-object tracking to multi-ob...
research
03/26/2023

SDTracker: Synthetic Data Based Multi-Object Tracking

We present SDTracker, a method that harnesses the potential of synthetic...
research
02/24/2022

GIAOTracker: A comprehensive framework for MCMOT with global information and optimizing strategies in VisDrone 2021

In recent years, algorithms for multiple object tracking tasks have bene...
research
09/20/2016

Adaptive Decontamination of the Training Set: A Unified Formulation for Discriminative Visual Tracking

Tracking-by-detection methods have demonstrated competitive performance ...
research
11/27/2018

Eliminating Exposure Bias and Loss-Evaluation Mismatch in Multiple Object Tracking

Identity Switching remains one of the main difficulties Multiple Object ...
research
11/07/2019

Improving Human Annotation in Single Object Tracking

Human annotation is always considered as ground truth in video object tr...
research
11/22/2022

β-Multivariational Autoencoder for Entangled Representation Learning in Video Frames

It is crucial to choose actions from an appropriate distribution while l...

Please sign up or login with your details

Forgot password? Click here to reset