Learning Accurate Entropy Model with Global Reference for Image Compression

Authors: Yichen Qian, Zhiyu Tan, Xiuyu Sun, Ming Lin, Dongyang Li, Zhenhong Sun, Li Hao, Rong Jin

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate that the proposed model outperforms the rate-distortion performance of most of the state-of-the-art methods in the industry.
Researcher Affiliation Industry Yichen Qian Zhiyu Tan Xiuyu Sun Alibaba Group Alibaba Group Alibaba Group Hangzhou, China Hangzhou, China Hangzhou, China yichen.qyc@alibaba-inc.com zhiyu.tzy@alibaba-inc.com xiuyu.sxy@alibaba-inc.com Ming Lin Dongyang Li Zhenhong Sun Alibaba Group Alibaba Group Alibaba Group Bellevue, WA, 98004, USA Hangzhou, China Hangzhou, China ming.l@alibaba-inc.com yingtian.ldy@alibaba-inc.com zhenhong.szh@alibaba-inc.com Hao Li Rong Jin Alibaba Group Alibaba Group Hangzhou, China Bellevue, WA, 98004, USA lihao.lh@alibaba-inc.com jinrong.jr@alibaba-inc.com
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any specific links to open-source code or explicitly state that the code is publicly available.
Open Datasets Yes The models were trained on color PNG images from CVPR workshop CLIC training dataset (http://challenge.compression.cc/). The models were optimized using Adam (Kingma & Ba, 2014) with a batch size of 8 and a patch size of 512 512 randomly extracted from the training dataset. Note that large patch size is necessary for the training of the reference model.
Dataset Splits No The paper mentions using the 'CLIC Validation dataset' for evaluation, but it does not specify a dataset split (e.g., percentages or counts) for training, validation, and test sets from a single dataset. It treats the CLIC training dataset and CLIC validation dataset as distinct sets for different purposes.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions 'Adam (Kingma & Ba, 2014)' as the optimizer but does not specify any software libraries or frameworks with version numbers (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes The models were optimized using Adam (Kingma & Ba, 2014) with a batch size of 8 and a patch size of 512 512 randomly extracted from the training dataset. Note that large patch size is necessary for the training of the reference model. As our combined entropy model have three predictive Gaussian parameters, we first trained the three modules with weight of 0.3 : 0.3 : 0.4 as a warm-up with 1000 epochs. After that, we trained three modules with weight of 0.1 : 0.1 : 0.8 because the third output is used for entropy coding in practice. In the experiments, we trained different models with different λ to evaluate the rate-distortion performance for various ranges of bit-rate.