Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Prune and Repaint: Content-Aware Image Retargeting for any Ratio

Authors: Feihong Shen, Chao Li, Yifeng Geng, Yongjian Deng, Hao Chen

NeurIPS 2024 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on the public Retarget Me benchmark and demonstrate through objective experimental results and subjective user studies that our method outperforms previous approaches in terms of preserving semantics and aesthetics, as well as better generalization across diverse aspect ratios.
Researcher Affiliation Collaboration Feihong Shen1,2,4 EMAIL Chao Li2 EMAIL Yifeng Geng2 EMAIL Yongjian Deng3 EMAIL Hao Chen 1,4 EMAIL 1Southeast University 2Alibaba Group 3Beijing University of Technology 4Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications (Southeast University), Ministry of Education, China
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes Codes will be available at https://github.com/fhshen2022/Prune Repaint.
Open Datasets Yes We evaluate the proposed method on the public image retargeting datasets, Retarget Me [27], which contains 80 images from various scenes. [27] M. Rubinstein, D. Gutierrez, O. Sorkine-Hornung, and A. Shamir. A comparative study of image retargeting. ACMSIGGRAPH, 2010.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning into train/validation/test sets.
Hardware Specification Yes Our method is implemented using Pytorch on a RTX 3090.
Software Dependencies Yes For the image-to-image repainting model in AR, we employ a composition of SD1.5 , Control Net-Inpainting and IP-Adapter [36].
Experiment Setup Yes The length of the sliding window in Section 3.3.1 is set to l = 25, and the threshold is set to η = 15.