Deep Supervision with Shape Concepts for Occlusion-Aware 3D Object Parsing
arXiv:1612.02699
Abstract
Monocular 3D object parsing is highly desirable in various scenarios including occlusion reasoning and holistic scene interpretation. We present a deep convolutional neural network (CNN) architecture to localize semantic parts in 2D image and 3D space while inferring their visibility states, given a single RGB image. Our key insight is to exploit domain knowledge to regularize the network by deeply supervising its hidden layers, in order to sequentially infer intermediate concepts associated with the final task. To acquire training data in desired quantities with ground truth 3D shape and relevant concepts, we render 3D object CAD models to generate large-scale synthetic data and simulate challenging occlusion configurations between objects. We train the network only on synthetic data and demonstrate state-of-the-art performances on real image benchmarks including an extended version of KITTI, PASCAL VOC, PASCAL3D+ and IKEA for 2D and 3D keypoint localization and instance segmentation. The empirical results substantiate the utility of our deep supervision scheme by demonstrating effective transfer of knowledge from synthetic data to real images, resulting in less overfitting compared to standard end-to-end training.
Accepted in CVPR 2017
References in corpus (12)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Caffe: Convolutional Architecture for Fast Feature Embedding
- Deep Convolutional Inverse Graphics Network
- Virtual Worlds as Proxy for Multi-Object Tracking Analysis
- Playing for Data: Ground Truth from Computer Games
- Render for CNN: Viewpoint Estimation in Images Using CNNs Trained with Rendered 3D Model Views
- Unsupervised Learning of 3D Structure from Images
- Towards Scene Understanding with Detailed 3D Object Representations
- WarpNet: Weakly Supervised Matching for Single-view Reconstruction
- Inferring 3D Object Pose in RGB-D Images
- Viewpoints and Keypoints
- Learning Dense Correspondence via 3D-guided Cycle Consistency