Structure-Consistent Weakly Supervised Salient Object Detection with Local Saliency Coherence

Authors: Siyue Yu, Bingfeng Zhang, Jimin Xiao, Eng Gee Lim3234-3242

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments show that our method achieves a new state-of-the-art performance on six benchmarks (e.g. for the ECSSD dataset: Fβ = 0.8995, Eξ = 0.9079 and MAE = 0.0489), with an average gain of 4.60% for F-measure, 2.05% for E-measure and 1.88% for MAE over the previous best method on this task.
Researcher Affiliation Academia Siyue Yu, Bingfeng Zhang, Jimin Xiao*, Eng Gee Lim School of Advanced Technology, Xi an Jiaotong-Liverpool University, Suzhou, China {siyue.yu, bingfeng.zhang, jimin.xiao, enggee.lim}@xjtlu.edu.cn
Pseudocode No The paper describes the proposed methods in text and mathematical formulas but does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code Yes Source code is available at http://github.com/siyueyu/SCWSSOD.
Open Datasets Yes We train our network on scribble annotated dataset S-DUTS (Zhang et al. 2020b) and evaluate our model on six widely-used salient object detection benchmarks: (1) ECSSD (Yan et al. 2013); (2) DUTOMRON (Yang et al. 2013); (3) PASCAL-S (Li et al. 2014); (4) HKU-IS (Li and Yu 2015); (5) THUR (Cheng et al. 2014); (6) DUTS-TEST (Wang et al. 2017).
Dataset Splits No The paper mentions training on S-DUTS and evaluating on six benchmarks. It details training parameters like epochs and image resizing but does not specify explicit training/validation/test dataset splits (e.g., percentages or sample counts for each split, or reference to standard splits for these roles).
Hardware Specification Yes All experiments are run on NVIDIA Ge Force RTX 2080 Ti.
Software Dependencies No The paper mentions using ResNet-50 as a backbone and GCPANet as a baseline, but does not provide specific version numbers for software libraries or frameworks (e.g., PyTorch, TensorFlow, CUDA).
Experiment Setup Yes The model is optimized by SGD with batch size of 16, momentum of 0.9 and weight decay of 5 * 10^-4. Additionally, we use triangular warm-up and decay strategies with the maximum learning rate of 0.01 and the minimum learning rate of 1 * 10^-5 to train the network with 40 epochs. During training, each image is resized to 320 * 320 with random horizontal flipping and random cropping. ... w, σP , and σI in Eq. (4) are set to 1, 6 and 0.1, respectively. β in Eq. (8) and Eq. (9) is set to 0.3.