Region-Adaptive Dense Network for Efficient Motion Deblurring

Authors: Kuldeep Purohit, A. N. Rajagopalan11882-11889

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive comparisons with prior art on benchmark dynamic scene deblurring datasets clearly demonstrate the superiority of the proposed networks via significant improvements in accuracy and speed, enabling almost real-time deblurring.
Researcher Affiliation Academia Kuldeep Purohit, A. N. Rajagopalan Indian Institute of Technology Madras, India kuldeeppurohit3@gmail.com, raju@ee.iitm.ac.in
Pseudocode No The paper provides architectural diagrams and descriptions but does not include pseudocode or explicit algorithm blocks.
Open Source Code No The paper does not include any explicit statements or links indicating the availability of open-source code for the described methodology.
Open Datasets Yes we perform training and evaluation of our network on the dynamic scene deblurring benchmark (Nah, Kim, and Lee 2017)
Dataset Splits Yes Following the same train-test split as in (Nah, Kim, and Lee 2017), we use 2103 pairs for training and 1111 pairs for evaluation.
Hardware Specification Yes We conduct our experiments on a PC with Intel Xeon E5 CPU, 256 GB RAM and an NVIDIA Titan X GPU.
Software Dependencies No The paper mentions optimizers and frameworks indirectly through citations, but does not provide specific version numbers for any software dependencies.
Experiment Setup Yes Training is done for 1 × 10^6 iterations using Adam optimizer with learning rate 0.0001 on patches of 256 × 256 and batch-size of 16.