Author name cluster

QiCheng Wang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers

2 author rows

EAAI Journal 2026 Journal Article

Diffusion-based lossy geometry compression for three dimensional point clouds

Haibo Zhang
Haoran Sun
Qicheng Wang
Xiao Cai
Xiao Wu
Mingquan Zhou
Guohua Geng

Point Clouds (PCs) serve as a crucial data representation for three dimensional (3D) spatial information. However, voxelizing Point Cloud (PC) models can lead to the loss of geometric structure information and an increase in computational complexity. Therefore, directly compressing the points of the PC model is often a more effective compression strategy. This paper proposes a diffusion-based lossy Point Cloud Compression (PCC) network that directly compresses the points of the PC model, thereby avoiding the information loss caused by voxelization and improving the quality of the reconstructed model. The proposed compression network is based on an AutoEncoder (AE) architecture, incorporating a multi-scale feature extraction module during the compression stage. The first channel of the module combines PointNet and Inception-ResNet (IRN) modules to extract more compact and efficient PC features, while the second one employs Sparse Convolution to optimize the storage representation, effectively capturing both geometric and semantic information. This process is essential for subsequent decompression and reconstruction phases. During the decoding stage, a diffusion-based generator utilizes shape latent variables as prior conditions to stochastically denoise the PC, generating a higher-quality reconstructed model. To further enhance the quality of the reconstructed surface model, the Chamfer Distance (CD) is integrated into the loss function to measure the geometric similarity between the original and reconstructed PC models. Experiments are conducted using ten models obtained from three different datasets. Experimental results demonstrate that the proposed method outperforms the latest Geometry-Point Cloud Compression (G-PCC) in terms of objective evaluation metrics, and achieves lower geometric information loss compared to other point-based compression networks.

Details DOI

IJCAI Conference 2025 Conference Paper

Stochasticity-aware No-Reference Point Cloud Quality Assessment

Songlin Fan
Wei Gao
Zhineng Chen
Ge Li
Guoqing Liu
Qicheng Wang

The evolution of point cloud processing algorithms necessitates an accurate assessment for their quality. Previous works consistently regard point cloud quality assessment (PCQA) as a MOS regression problem and devise a deterministic mapping, ignoring the stochasticity in generating MOS from subjective tests. This work presents the first probabilistic architecture for no-reference PCQA, motivated by the labeling process of existing datasets. The proposed method can model the quality judging stochasticity of subjects through a tailored conditional variational autoencoder (CVAE) and produces multiple intermediate quality ratings. These intermediate ratings simulate the judgments from different subjects and are then integrated into an accurate quality prediction, mimicking the generation process of a ground truth MOS. Specifically, our method incorporates a Prior Module, a Posterior Module, and a Quality Rating Generator, where the former two modules are introduced to model the judging stochasticity in subjective tests, while the latter is developed to generate diverse quality ratings. Extensive experiments indicate that our approach outperforms previous cutting-edge methods by a large margin and exhibits gratifying crossdataset robustness. Codes are available at https: //git. openi. org. cn/OpenPointCloud/nrpcqa.

PDF Details DOI

ICML Conference 2024 Conference Paper

Prompt-based Visual Alignment for Zero-shot Policy Transfer

Haihan Gao
Rui Zhang 0040
Qi Yi
Hantao Yao
Haochen Li 0002
Jiaming Guo
Shaohui Peng
Yunkai Gao 0001

Overfitting in RL has become one of the main obstacles to applications in reinforcement learning(RL). Existing methods do not provide explicit semantic constrain for the feature extractor, hindering the agent from learning a unified cross-domain representation and resulting in performance degradation on unseen domains. Besides, abundant data from multiple domains are needed. To address these issues, in this work, we propose prompt-based visual alignment (PVA), a robust framework to mitigate the detrimental domain bias in the image for zero-shot policy transfer. Inspired that Visual-Language Model (VLM) can serve as a bridge to connect both text space and image space, we leverage the semantic information contained in a text sequence as an explicit constraint to train a visual aligner. Thus, the visual aligner can map images from multiple domains to a unified domain and achieve good generalization performance. To better depict semantic information, prompt tuning is applied to learn a sequence of learnable tokens. With explicit constraints of semantic information, PVA can learn unified cross-domain representation under limited access to cross-domain data and achieves great zero-shot generalization ability in unseen domains. We verify PVA on a vision-based autonomous driving task with CARLA simulator. Experiments show that the agent generalizes well on unseen domains under limited access to multi-domain data.

Details