First Step toward Model-Free, Anonymous Object Tracking with Recurrent Neural Networks

Abstract

In this paper, we propose and study a novel visual object tracking approach based on convolutional networks and recurrent networks. The proposed approach is distinct from the existing approaches to visual object tracking, such as filtering-based ones and tracking-by-detection ones, in the sense that the tracking system is explicitly trained off-line to track anonymous objects in a noisy environment. The proposed visual tracking model is end-to-end trainable, minimizing any adversarial effect from mismatches in object representation and between the true underlying dynamics and learning dynamics. We empirically show that the proposed tracking approach works well in various scenarios by generating artificial video sequences with varying conditions; the number of objects, amount of noise and the match between the training shapes and test shapes.

References in corpus (4)

• Recurrent Models of Visual Attention
• DRAW: A Recurrent Neural Network For Image Generation
• Multiple Object Recognition with Visual Attention
• Attentional Neural Network: Feature Selection Using Cognitive Feedback

Cited by in corpus (11)

• Good Features to Correlate for Visual Tracking
• Deep Reinforcement Learning for Visual Object Tracking in Videos
• Deep 6-DOF Tracking
• Spatially Supervised Recurrent Convolutional Neural Networks for Visual Object Tracking
• Similarity Mapping with Enhanced Siamese Network for Multi-Object Tracking
• Fully-Convolutional Siamese Networks for Object Tracking
• Recurrent Filter Learning for Visual Tracking
• Adversarial-based neural networks for affect estimations in the wild
• Object Tracking Using Spatio-Temporal Future Prediction
• Pixel-wise object tracking
• Object-Adaptive LSTM Network for Real-time Visual Tracking with Adversarial Data Augmentation