DME: Unveiling the Bias for Better Generalized Monocular Depth Estimation

Authors: Songsong Yu, Yifan Wang, Yunzhi Zhuge, Lijun Wang, Huchuan Lu

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments show that DME achieves state-of-the-art performance on both NYU-Depth v2 and KITTI, and also delivers favorable zero-shot generalization capability on unseen datasets.
Researcher Affiliation Academia Songsong Yu, Yifan Wang, Yunzhi Zhuge, Lijun Wang*, Huchuan Lu Dalian University of Technology 22209083@mail.dlut.edu.cn, {wyfan, ljwang, lhchuan}@dlut.edu.cn, zgyzzgyz@gmail.com
Pseudocode No The paper describes its methods in prose and through diagrams (Figure 3), but does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code Yes The code can be obtained at https://github.com/YUsong360/DME-Unveilingthe-bias.
Open Datasets Yes Our training data consists of NYUD v2 (Silberman et al. 2012) and KITTI (Geiger et al. 2013) training datasets.
Dataset Splits No The paper mentions using NYUD v2 and KITTI datasets for training and testing, but does not explicitly state the training/validation/test splits (e.g., percentages, sample counts, or specific predefined split files) for reproducibility.
Hardware Specification Yes The training process is performed on one NVIDIA Ge Force RTX 3090Ti GPU, taking about 20 hours in total.
Software Dependencies No The paper mentions the use of the Adam optimizer but does not specify version numbers for any software, libraries, or frameworks used (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes During training, the Adam optimizer (Kingma and Ba 2014) is employed with a batch size of 2 and a weight decay of 1e-2. The initial learning rate is set to 3e-4.