Learning to Appreciate the Aesthetic Effects of Clothing
Authors: Jia Jia, Jie Huang, Guangyao Shen, Tao He, Zhiyuan Liu, Huanbo Luan, Chao Yan
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Employing upper-body menswear images downloaded from several global online clothing shops as experimental data, the results indicate that the proposed three-level framework can help to capture the subtle relationship between visual features and aesthetic words better compared to several baselines. In this section, we first conduct several objective experiments to validate the SDAE-GCL by evaluating the mapping effects between visual features of clothing images and coordinate values in the image-scale space. Then we show the effectiveness of the proposed VF-ISS-AWS framework through some interesting demonstrations. |
| Researcher Affiliation | Collaboration | Jia Jia1, Jie Huang1, Guangyao Shen1, Tao He2, Zhiyuan Liu1 , Huanbo Luan1 and Chao Yan3 1Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China Tsinghua National Laboratory for Information Science and Technology (TNList) Key Laboratory of Pervasive Computing, Ministry of Education 2School of Computer Science, Sichuan University, Chengdu 610065 3Beijing Samsung Telecom R& D center |
| Pseudocode | Yes | Algorithm 1 Stacked Denoising Autoencoder Guided by Correlative Labels |
| Open Source Code | No | The paper does not contain any explicit statement about releasing source code or a link to a code repository. |
| Open Datasets | No | D1: Labeled Dataset. This dataset contains 5500 images downloaded from Amazon... D2: Unlabeled Dataset. In order to make our model more applicable in different data sources, we select another online shopping website JD as the data source and fetch 130,316 images... |
| Dataset Splits | Yes | All the experiments are performed on five-folder cross-validation. |
| Hardware Specification | Yes | on this condition the experiment lasts for about two hours on an environment with dual-core 2.10GHZ CPU, 64GB memory. |
| Software Dependencies | No | The paper mentions software tools like Word Net and algorithms like SVM, but does not provide specific version numbers for any software dependencies. |
| Experiment Setup | No | The paper mentions hyper-parameters like α, λ1, λ2, λ3, and β, but does not provide their specific concrete values. It states that "we take 5 layers in our experiments" for the hidden layer number, but lacks other specific setup details. |