End-to-End United Video Dehazing and Detection

Authors: Boyi Li, Xiulian Peng, Zhangyang Wang, Jizheng Xu, Dan Feng

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental A thorough study has been conducted over a number of structure options, to identify the best temporal fusion strategy. Experiment Results on Video Dehazing. Table 1: PSNR/SSIM Comparisons of Various Structures.
Researcher Affiliation Collaboration 1Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology 2Microsoft Research, Beijing, China 3Department of Computer Science and Engineering, Texas A&M University
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide an explicit statement about releasing source code or a direct link to a code repository for the methodology described.
Open Datasets Yes We created a synthetic hazy video dataset based on (1), using 18 videos selected from the TUM RGB-D Dataset (Sturm et al. 2012) and We synthesize hazy videos with various haze levels for a subset of ILSVRC2015 VID dataset (Russakovsky et al. 2015)
Dataset Splits No The paper explicitly states training and testing set sizes and compositions (e.g., 'training set, consisting of 5 videos with 100,000 frames' and 'testing set called Test Set V1, consisting of the rest 13 relatively short video clips with a total of 403 frames'), but it does not specify a distinct validation set or its size/split.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with versions like Python 3.8, CPLEX 12.4) needed to replicate the experiment.
Experiment Setup Yes When training EVD-Net, the momentum and the decay parameters are set to 0.9 and 0.0001, respectively, with a batch size of 8. We adopt the Mean Square Error (MSE) loss, which has been shown in (Li et al. 2017a; 2017b) that it is well aligned with SSIM and visual quality. we first tune only the fully-connected layers in the high-level detection part of EVD-Net for 90,000 iterations, and then tune the entire concatenated pipeline for another 10,000 iterations.