FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection

Rukhovich, Danila; Vorontsova, Anna; Konushin, Anton

doi:10.1007/978-3-031-20080-9_28

Cited by 63 publications

(30 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In traditional navigation, obstacles can be accurately located but not classified by LiDAR. In this study, we use the fully convolution anchor-free 3D object detection (FCAF3D) (Rukhovich et al , 2022), a 3D object detection network with a good balance between speed and accuracy, to detect obstacles. Different from the original text, we extend it to detect dynamic targets, such as pedestrians, by adding categories.…”

Section: Methodsmentioning

confidence: 99%

“…Our software system is based on an open-source ROS system, and all modules communicate with each other through a topic. Open-source FCAF3D (Rukhovich et al , 2022), with excellent speed and accuracy, is used for the 3D target detector, which is a SOFA network for indoor 3D target detection. The informed-RRT* (Gammell et al , 2014) algorithm is used to carry out path planning, and the detailed process can be found in Section 3.5.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A hybrid obstacle avoidance method for mobile robot navigation in unstructured environment

Zhou

Feng²,

Chou³

2022

View full text Add to dashboard Cite

Purpose Wheeled mobile robots (WMR) are the most widely used robots. Avoiding obstacles in unstructured environments, especially dynamic obstacles such as pedestrians, is a serious challenge for WMR. This paper aims to present a hybrid obstacle avoidance method that combines an informed-rapidly exploring random tree* algorithm with a three-dimensional (3D)-object detection approach and model prediction controller (MPC) to conduct obstacle perception, collision-free path planning and obstacle avoidance for WMR in unstructured environments. Design/methodology/approach Given a reference orientation and speed, the hybrid method uses parametric ellipses to represent obstacle expansion boundaries based on the 3D target detection results, and a collision-free reference path is planned. Then, the authors build on a model predictive control for tracking the collision-free reference path by incorporating the distance between the robot and obstacles. The proposed framework is a mapless method for WMR. Findings The authors present experimental results with a mobile robot for obstacle avoidance in indoor environments crowded with obstacles, such as chairs and pedestrians. The results show that the proposed hybrid obstacle avoidance method can satisfy the application requirements of mobile robots in unstructured environments. Originality/value In this study, the parameter ellipse is used to represent the area occupied by the obstacle, which takes the velocity as the parameter. Therefore, the motion direction and position of dynamic obstacles can be considered in the planning stage, which enhances the success rate of obstacle avoidance. In addition, the distance between the obstacle and robot is increased in the MPC optimization function to ensure a safe distance between the robot and the obstacle.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

A hybrid obstacle avoidance method for mobile robot navigation in unstructured environment

Zhou

Feng²,

Chou³

2022

View full text Add to dashboard Cite

show abstract

“…For the assignment strategy, we adopt the assignment strategy in ref. [10]. We set the threshold

{N_{\text{loc}}}

for the feature layer selection.…”

Section: Assignment Strategymentioning

confidence: 99%

“…GSDN [9] is an anchor‐based method consisting of full convolution, and consists of two parts: encoder and decoder. FCAF3D [10] is an anchor‐free method consisting of full convolution, and oriented‐bounding‐box parameters were obtained by data‐driven rather than using prior parameters. The convolution in GSDN and FCAF3D are sparse, and the memory efficiency of these two methods are higher than voting‐based and transformer‐based method.…”

Section: Introductionmentioning

confidence: 99%

BMFN3D: Bidirectional multilayer fusion network for indoor 3D object detection

Cheng

Zhang

2022

Electronics Letters

View full text Add to dashboard Cite

Among various indoor 3D object detection methods, one of which is based on 3D convolution. However, in current 3D convolutional method for indoor 3D object detection, feature fusion is only implemented in a single direction and adjacent layer, which leads to the feature loss during feature propagation and weakens the detection performance. To solve the above problem, a bidirectional multilayer fusion network (BMFN) is proposed to enrich features for prediction. Furthermore, we propose a new network based on BMFN for 3D object detection, BMFN3D. In BMFN3D, the channel and spatial attention mechanisms are introduced into the backbone and head of detection network, so as to enhance the selection of effective feature information. Finally, a new calculation method for regression loss is proposed with the consideration of the 3D spatial information. We validate our method with two indoor datasets of SUN RGB-D and ScanNet, and achieve state-of-the-art results.

show abstract

“…Compared to other methods, 3D sparse convolutional methods are memory-efficient and scale to large scenes well without sacrificing point density. Up until very recently, such methods lacked accuracy, yet due to the recent advances in the field, fast and scalable yet accurate methods were developed [1].…”

Section: Introductionmentioning

confidence: 99%

TR3D: Towards Real-Time Indoor 3D Object Detection

Rukhovich¹,

Воронцова²,

Konushin³

2023

Preprint

View full text Add to dashboard Cite

Recently, sparse 3D convolutions have changed 3D object detection. Performing on par with the voting-based approaches, 3D CNNs are memory-efficient and scale to large scenes better. However, there is still room for improvement. With a conscious, practice-oriented approach to problem-solving, we analyze the performance of such methods and localize the weaknesses. Applying modifications that resolve the found issues one by one, we end up with TR3D: a fast fully-convolutional 3D object detection model trained end-to-end, that achieves state-of-the-art results on the standard benchmarks, ScanNet v2, SUN RGB-D, and S3DIS. Moreover, to take advantage of both point cloud and RGB inputs, we introduce an early fusion of 2D and 3D features. We employ our fusion module to make conventional 3D object detection methods multimodal and demonstrate an impressive boost in performance. Our model with early feature fusion, which we refer to as TR3D+FF, outperforms existing 3D object detection approaches on the SUN RGB-D dataset. Overall, besides being accurate, both TR3D and TR3D+FF models are lightweight, memory-efficient, and fast, thereby marking another milestone on the way toward real-time 3D object detection. Code is available at https://github.com/SamsungLabs/tr3d.

show abstract

FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection

Cited by 63 publications

References 31 publications

A hybrid obstacle avoidance method for mobile robot navigation in unstructured environment

A hybrid obstacle avoidance method for mobile robot navigation in unstructured environment

BMFN3D: Bidirectional multilayer fusion network for indoor 3D object detection

TR3D: Towards Real-Time Indoor 3D Object Detection

Contact Info

Product

Resources

About