CV-Zoo

3D

• OctNet: Learning Deep 3D Representations at High Resolutions
• Unsupervised Learning of Long-Term Motion Dynamics for Videos

Scene Related

• Scene Graph Generation by Iterative Message Passing

Discriptive Image Related

• A Hierarchical Approach for Generating Descriptive Image Paragraphs

Reinforcement Learning Related

• Target-driven Visual Navigation in Indoor Scenes using Deep Reinforcement Learning

Ideas:

1. 3D+VQA
2. GAN+3D or VQA

Datasets:

1. CLEVR: Compositional Language and Elementary Visual Reasoning