ACR-Pose: Adversarial Canonical Representation Reconstruction Network for Category Level 6D Object Pose Estimation

Fan, Zhaoxin; Song, Zhengbo; Xu, Jian; Wang, Zhicheng; Wu, Kejian; Liu, Hongyan; He, Jun

doi:10.48550/arxiv.2111.10524

Cited by 5 publications

(6 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To validate the effectiveness of AG-Net, we conduct comparative experiments on both the CAMERA25 and REAL275 datasets with existing methods including those that utilize prior information [17,32,33,[37][38][39][40][41][42][43][44] and those that do not utilize prior information [8,9,13,31,34,45,46].…”

Section: Comparisons With Existing Methodsmentioning

confidence: 99%

AG-Net: Category-level 6D Pose Estimation Network Based on Attention Mechanism and Global Enhancement

Bowen,

Jinlong,

Suqin

et al. 2024

Preprint

View full text Add to dashboard Cite

How to effectively enhance feature extraction is a challenge faced by current 6D pose estimation methods. To address this issue, we propose a novel 6D pose estimation network based on VI-Net, shortened as AG-Net, which uses ECA block and Global enhancev module to enhance feature extraction: ECA block embeds a channel attention mechanism into the convolutional layers, replacing the fully connected layers with 1×1Conv to capture relationships between different channels and improves the performance of feature extraction. The Global enhancev module further processes the received information and enhances feature extraction by fusing global features, effectively balancing performance and speed, and better estimating the translation and size of objects. We applied the proposed AG-Net to category-level 6D pose estimation tasks and tested it on the REAL275 and CAMER25 datasets using IOU 3D intersection and n◦mcm evaluation metrics. The results showed that AG-Net outperformed current state-of-the-art methods. Our code and models are available at https://github.com/AFESDTTM/AG-Net

show abstract

Section: Comparisons With Existing Methodsmentioning

confidence: 99%

AG-Net: Category-level 6D Pose Estimation Network Based on Attention Mechanism and Global Enhancement

Bowen,

Jinlong,

Suqin

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Kai et al [3] utilize a transformer network [8] to model the global structure similarity between prior and target object, based on which the object semantic information is injected into the prior feature to dynamically adapt the category-level prior to each particular object. Fan et al [9] adopt a shape prior guided reconstruction network and a discriminator network to learn high-quality canonical representations. Zhang et al [40] use the shape priors as the indicator to predict pose and zero-mean residual vectors which encapsulate the spatial cues of the pose and enable geometry-guided consistency terms.…”

Section: Prior-based Methodsmentioning

confidence: 99%

“…To overcome intra-class variation, the prior deformation, as a practical module, has been widely adopted by recent works [9,3,21]. The vanilla version of prior deformation can be divided into two parts: 1) generating shape priors and 2) leveraging shape priors to develop prior deformation techniques.…”

Section: Preliminarymentioning

confidence: 99%

Prior-free Category-level Pose Estimation with Implicit Space Transformation

Liu¹,

Chen²,

Ye³

et al. 2023

Preprint

View full text Add to dashboard Cite

Category-level 6D pose estimation aims to predict the poses and sizes of unseen objects from a specific category. Thanks to prior deformation, which explicitly adapts a category-specific 3D prior (i.e., a 3D template) to a given object instance, prior-based methods attained great success and have become a major research stream. However, obtaining category-specific priors requires collecting a large amount of 3D models, which is labor-consuming and often not accessible in practice. This motivates us to investigate whether priors are necessary to make prior-based methods effective. Our empirical study shows that the 3D prior itself is not the credit to the high performance. The keypoint actually is the explicit deformation process, which aligns camera and world coordinates supervised by worldspace 3D models (also called canonical space). Inspired by these observation, we introduce a simple prior-free implicit space transformation network, namely IST-Net, to transform camera-space features to world-space counterparts and build correspondence between them in an implicit manner without relying on 3D priors. Besides, we design camera-and world-space enhancers to enrich the features with pose-sensitive information and geometrical constraints, respectively. Albeit simple, IST-Net becomes the first prior-free method that achieves state-of-the-art performance, with top inference speed on the REAL275 dataset. Our code and models will be publicly available.

show abstract

“…A category-level shape prior is found to be beneficial for pose estimation accuracy in [29] and further improved in [30], [31], [32]. DualPoseNet [33], 6D-ViT [34], ACR-Pose [35], and CPPF [36] proposed to incorporate rotation-invariant embedding, Transformer networks, Generative Adversarial Networks, and deep pointpair-feature, respectively. However, these techniques require high-quality depth input provided by opaque objects with Lambertian light reflectance.…”

Section: B Opaque Object Category-level Pose Estimationmentioning

confidence: 99%