Learning deep structured network for weakly supervised change detection

Authors: Salman Khan, Xuming He, Fatih Porikli, Mohammed Bennamoun, Ferdous Sohel, Roberto Togneri

IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our evaluations on four benchmark datasets demonstrate superior detection and localization performance. [...] 4 Experiments [...] Our experimental results demonstrate that the proposed approach outperforms the state-of-the-art by a large margin (Sec. 4.3). The key contributions of our work include: [...] Furthermore, we perform a rigorous evaluation on three other relevant datasets.
Researcher Affiliation Academia Salman Khan1,2 , Xuming He3 , Fatih Porikli2, Mohammed Bennamoun4 Ferdous Sohel5 and Roberto Togneri4 1Data61-CSIRO, Canberra, Australia 2Australian National University, Canberra, Australia 3Shanghai Tech University, Shanghai, China 4The University of Western Australia, Perth, Australia 5Murdoch University, Perth, Australia
Pseudocode No The paper describes algorithms but does not provide pseudocode or a clearly labeled algorithm block.
Open Source Code No The paper does not contain an explicit statement about releasing its source code or a link to a code repository for the described methodology.
Open Datasets Yes We evaluate our method on the following four datasets. [...] CDnet 2014 Dataset: [Wang et al., 2014] [...] AICD 2012 Dataset: Aerial Image Change Detection (AICD) dataset [Bourdis et al., 2011] [...] GASI 2015 Dataset: Geoscience Australia Satellite Image (GASI) dataset is a custom built dataset based on the changes occurred during 1999 2015 in a 100 100 km2 area in the east of city of Melbourne in Victoria, Australia [Khan et al., 2017]. [...] PCD 2015 Dataset: Panoramic change detection dataset [Sakurada and Okatani, 2015]
Dataset Splits Yes For each test image pair, we find K closely matching pairs from the training set using a KNN search and average their foreground label proportion to estimate τ (details in Sec. 4.3). [...] Note that for the training, we validate a fixed-value τ, which is faster than using KNN. As the pixel-level labels are unavailable in training, we set τ to a value which gives coverage of at least 15% of each image on a validation set.
Hardware Specification No The paper does not explicitly describe the hardware specifications (e.g., specific GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions software like 'lib-linear package' and 'Caffe', and refers to network architectures like 'VGGnet' and 'FCN-VGG16-16s', but it does not specify version numbers for these or other software dependencies.
Experiment Setup Yes The network weights are initialized from a pre-trained VGG network (on Image Net). [...] enlarged paired images of size 512x512 are fed to the CNN. Moreover, the convolution filter size in FC1 (segmentation branch) is kept to 1x1 (in contrast to a 7x7 filter size in FC1 ) to avoid the additional decrease in resolution of the 32x32 output map. [...] As the pixel-level labels are unavailable in training, we set τ to a value which gives coverage of at least 15% of each image on a validation set.