TextManiA: Enriching Visual Feature by

Text-driven Manifold Augmentation

Moon Ye-Bin1, Jisoo Kim2, Hongyeob Kim3, Kilho Son4, Tae-Hyun Oh1
1POSTECH, 2Columbia University, 3Sungkyunkwan University, 4Microsoft Azure
International Conference on Computer Vision (ICCV) 2023
Paper arXiv Video Poster Code
Interpolation end reference image.

TextManiA semantically enriches the target visual feature space by leveraging visually mimetic texts, encoded with general language models and transferred to the target visual feature space.

Abstract

Recent label mix-based augmentation methods have shown their effectiveness in generalization despite their simplicity, and their favorable effects are often attributed to semantic-level augmentation. However, we found that they are vulnerable to highly skewed class distribution, because scarce data classes are rarely sampled for inter-class perturbation.

We propose TextManiA, a text-driven manifold augmentation method that semantically enriches visual feature spaces, regardless of data distribution. TextManiA augments visual data with intra-class semantic perturbation by exploiting easy-to-understand visually mimetic words, i.e., attributes. To this end, we bridge between the text representation and a target visual feature space, and propose an efficient vector augmentation.

To empirically support the validity of our design, we devise two visualization-based analyses and show the plausibility of the bridge between two different modality spaces. Our experiments demonstrate that TextManiA is powerful in scarce samples with class imbalance as well as even distribution. We also show compatibility with the label mix-based approaches in evenly distributed scarce data.

Method


Interpolation end reference image.

TextManiA

  • Visual feature augmentation by conveying attribute information from the text to visual feature space
  • Intra-class semantic perturbation; densifying samples
  • Helpful in augmenting sparse samples in long-tail case
  • Complementary to other aug. in deficient data case
    •

Training Process

1. Given image $\mathbf{I}_0$ & label $T_0$
2. Synthesize text variant $T_1$
3. Compute difference vector $\mathbf{\Delta}_{0\to1}$
4. Add projected $\mathbf{\Delta}_{0\to1}$ with image feature $\mathbf{f}_{I_0}$
5. Train the model with augmented feature
     $\hat{\mathbf{f}}_{I_0} = \mathbf{f}_{I_0} + \alpha\cdot \mathtt{proj}(\mathbf{\Delta}_{0\to1})$

Characteristics of Difference Vector


t-SNE Plot

Interpolation end reference image.
  • From the same attr. → clustered regardless of the class
  • Slight difference → surrounding meaning left to some extent
    •

Image Manipulation

Interpolation end reference image.
  • Embed attribute while preserving semantics
  • Similar effects to img-level augmentation by text w/o complicated operation
    •

Acknowledgement

This work was partly supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No.2021-0-02068, Artificial Intelligence Innovation Hub), (No.2022-0-00124, Development of Artificial Intelligence Technology for Self-Improving Competency-Aware Learning Capabilities), and (No. 2020-0-00004, Development of Previsional Intelligence based on Long-term Visual Memory Network).

BibTeX

@inproceedings{yebin2023textmania,
  title     = {TextManiA: Enriching Visual Feature by Text-driven Manifold Augmentation},
  author    = {Moon Ye-Bin and Jisoo Kim and Hongyeob Kim and Kilho Son and Tae-Hyun Oh},
  booktitle = {ICCV},
  year      = {2023},
}