SceneNet: An annotated model generator for indoor scene understanding

By Ankur Handa, Viorica Patraucean, Simon Stent, Roberto Cipolla. 2016

111 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Geology, 3D Vision & Pose

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

1. Abhinav Kumar Gupta
2. Nathan Paczan
3. Mike Roberts
4. Ali Farhadi
5. Stefan Leutenegger
6. Sergey I. Nikolenko
7. Roozbeh Mottaghi
8. Andrew J. Davison

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Jitendra Malik
3. Abhinav Kumar Gupta
4. Dacheng Tao
5. Li Fei-Fei
6. Trevor Darrell
7. Andrea Vedaldi
8. Ming-Hsuan Yang

In Geology

1. Sergey I. Nikolenko
2. Stefan Leutenegger
3. Nathan Paczan
4. Mike Roberts
5. John McCormac
6. Simon Stent
7. Wenbin Li
8. Dimos Tzoumanikas

In 3D Vision & Pose

1. Sergey I. Nikolenko
2. Nathan Paczan
3. Mike Roberts
4. Abhinav Kumar Gupta
5. Stefan Leutenegger
6. Ali Farhadi
7. Yuke Zhu
8. Simon Stent

Citations in the corpus

• Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding (2020, methodology): “Non-photorealistic datasets and environments [29, 30, 70, 78] also play an important role in scene understanding research because they can be rendered very efficiently.”
• Target-driven visual navigation in indoor scenes using deep reinforcement learning (2016, methodology): “[30] propose a new Screenshots of our framework and other simulated learning frameworks: ALE [36], ViZDoom [37], UETorch [38], Project Malmo [39], SceneNet [40], TORCS [41], SYNTHIA [42], Virtual KITTI [43]. search algorithm based on the integration of Monte-Carlo tree search with deep RL that beats”
• Understanding RealWorld Indoor Scenes with Synthetic Data (2015, methodology): “In this work, we build upon SceneNet [19], a repository of annotated synthetic 3D scenes, to collect potentially unlimited labelled training data with perfect ground truth for per-pixel labelling.”
• Configurable 3D Scene Synthesis and 2D Image Rendering with Per-pixel Ground Truth Using Stochastic Grammars (2017, methodology): “timation [18, 27, 39, 104, 109, 111, 126, 139, 151], depth prediction [118], pedestrian detection [49, 81, 94, 127], action recognition [100, 101, 115], semantic segmentation [103], sceneunderstanding[45, 46, 60, 97],aswellasinbenchmark datasets [47]. Previously, synthetic imagery, generated on the ”
• Arena: A General Evaluation Platform and Building Toolkit for Multi-Agent Intelligence (2019, background): “…specific tasks, such as first person shooting [104], Real-Time Strategy (RTS) [93], vision understanding [76], in-door scene understanding [12, 15, 29, 80], and interaction [45, 81, 103], or specific games, such as driving [20, 105], handling blocks [53], Pac-Man [99], Starcraft [16, 19, 97, 102],…”
• Synthetic Data for Deep Learning (2019, methodology): “The dataset is actually an extension of SceneNet [232, 234], an annotated model generator for indoor scene understanding that can use existing datasets of 3D object models and place them in 3D environments with synthetic annotation.”
• InteriorNet: Mega-scale Multi-sensor Photo-realistic Indoor Scenes Dataset (2018, background): “In other papers [14], motion-capture trajectories have been captured to solve this – but such an approach is limited to small datasets where it is feasible to manually capture such trajectories.”
• HoME: a Household Multimodal Environment (2017, unknown): “These datasets include Stanford Scenes (1723 layouts) [34], Matterport3D [35] (90 layouts), sceneNN (100 layouts) [36], SceneNet (57 layouts) [37], and SceneNet RGB-D (57 layouts) [38].”