Robust Image Ordinal Regression with Controllable Image Generation
Authors: Yi Cheng, Haochao Ying, Renjun Hu, Jinhong Wang, Wenhao Zheng, Xiao Zhang, Danny Chen, Jian Wu
IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate the effectiveness of our new CIG approach in three different image ordinal regression scenarios. The results demonstrate that CIG can be flexibly integrated with off-the-shelf image encoders or ordinal regression models to achieve improvement, and further, the improvement is more significant for minority categories. We conduct extensive experiments on three highly different image ordinal regression scenarios (datasets), i.e., age estimation (Adience), diabetic retinopathy diagnosis (DR), and image quality ranking (Aesthetics), to evaluate the effectiveness of our CIG approach. |
| Researcher Affiliation | Collaboration | Yi Cheng1 , Haochao Ying2 , Renjun Hu3 , Jinhong Wang4 , Wenhao Zheng4 , Xiao Zhang5 , Danny Chen 6 and Jian Wu2,7 1School of Software Technology, Zhejiang Univerisity 2School of Public Health, Zhejiang University 3Alibaba Group 4College of Computer Science and Technology, Zhejiang University 5School of Computer Science and Technology, Shandong University 6Department of Computer Science and Engineering, University of Notre Dame 7Second Affiliated Hospital School of Medicine |
| Pseudocode | No | The paper describes the model architecture and training process in detail but does not include any explicit pseudocode blocks or algorithms. |
| Open Source Code | Yes | Our CIG is implemented using Py Torch [Paszke et al., 2019], which is available at Git Hub3. 3https://github.com/Ch3ng Y1/Controllable-Image-Generation |
| Open Datasets | Yes | We use three public datasets to evaluate our CIG. (1) Adience [Levi and Hassner, 2015] is a face image dataset from Flickr. (2) DR (Diabetic Retinopathy) [Liu et al., 2018a] contains high-resolution fundus images of patients. (3) Aesthetics [Schifanella et al., 2015] is another Flickr image dataset whose images are rated by the image quality. |
| Dataset Splits | Yes | We use 5-fold (on Adience and Aesthetics) or 10-fold (on DR) cross-validation, and report the average results. |
| Hardware Specification | Yes | All the experiments are conducted on a machine with 16 Intel(R) Xeon(R) Gold 6226R 2.90GHz CPUs and an NVIDIA RTX 3090 GPU. |
| Software Dependencies | No | The paper mentions "Py Torch [Paszke et al., 2019]" and "Adam optimizer", but it does not specify exact version numbers for PyTorch or any other software libraries, which is necessary for reproducibility. |
| Experiment Setup | Yes | We adopt the default Adam optimizer and a batch size of 18 for model training. The learning rates for the encoder and generator are set as 1e-4 and 5e-3, respectively. We optimize hyper-parameters on Adience with α {1, 2, 5}, β {1, 2, 5}, λ {0, 0.1, . . . , 1}, and τ {0.1, 0.2, . . . , 0.9}, and choose α = 5, β = 2, λ = 0.2, and τ = 0.2 for all our tests. |