IPDN: Image-enhanced Prompt Decoding Network for 3D Referring Expression Segmentation

Authors: Qi Chen, Changli Wu, Jiayi Ji, Yiwei Ma, Danni Yang, Xiaoshuai Sun

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Comprehensive experiments demonstrate that IPDN outperforms the state-of-the-art by 1.9 and 4.2 points in m Io U metrics on the 3D-RES and 3D-GRES tasks, respectively.
Researcher Affiliation Academia 1Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University, 361005, P.R. China. 2National University of Singapore. EMAIL, EMAIL, EMAIL, EMAIL, EMAIL, EMAIL
Pseudocode No The paper describes the methods using mathematical formulations and textual explanations but does not include explicit pseudocode or algorithm blocks.
Open Source Code Yes Code https://github.com/80chen86/IPDN
Open Datasets Yes We utilize the Scan Refer dataset (Chen, Chang, and Nießner 2020) to evaluate our method... We use the Multi3DRefer (Zhang, Gong, and Chang 2023) dataset to evaluate our model s performance on the 3D-GRES task...
Dataset Splits Yes We utilize the Scan Refer dataset (Chen, Chang, and Nießner 2020) to evaluate our method... We categorized object classes based on their frequency of appearance in the training set and conducted testing accordingly, as shown in Tab. 3.
Hardware Specification Yes All experiments are conducted using the Py Torch framework on an NVIDIA Ge Force RTX 3090 GPU.
Software Dependencies No The paper mentions 'PyTorch framework' but does not specify a version number or other software dependencies with their versions.
Experiment Setup Yes In our experiments, we apply the Poly RL strategy to adjust the learning rate starting from 0.0001, with a decay power of 4.0. The batch size is set to 16. The number of queries m is set to 128. The decoder consists of 6 layers. The hyperparameter k in sec.3.2 is set to 8, and the hyperparameter r in sec. 3.3 is 0.75. In the loss function, the weights λb, λp, and λc are set to 1.0, 0.1, and 0.1 respectively.