Towards Squeezing-Averse Virtual Try-On via Sequential Deformation

Authors: Sang-Heon Shim, Jiwoo Chung, Jae-Pil Heo

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results show that our SDVITON successfully resolves both types of artifacts and outperforms the baseline methods.
Researcher Affiliation Academia Sang-Heon Shim, Jiwoo Chung, Jae-Pil Heo* Sungkyunkwan University {ekzmwww, wldn0202, jaepilheo}@skku.edu
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes Source code will be available at https://github.com/SHShim0513/SD-VITON.
Open Datasets Yes We conduct experiments on a high-resolution virtual try-on dataset introduced by VITON-HD (Choi et al. 2021). It contains 11, 647 image pairs for the training phase and 2, 032 pairs for the evaluation, each of which has a frontview woman and a top clothes with 1024 768 resolution.
Dataset Splits No The paper states '11, 647 image pairs for the training phase and 2, 032 pairs for the evaluation', but does not explicitly mention a separate validation set or specific proportions for all three splits (train/validation/test).
Hardware Specification No The paper does not provide specific details regarding the hardware (e.g., GPU model, CPU type, memory) used for running the experiments.
Software Dependencies No The paper does not list specific version numbers for software dependencies (e.g., Python, PyTorch, CUDA versions) required for reproducibility.
Experiment Setup Yes LTOCG = λCELCE + λL1(LL1 + LM L1 ) +(LVGG + LM VGG) + Lc GAN + λTVLTV + Lz-dist, (13) where λCE, λL1, λTV is set to 10, 10, and 2 for a balance.