Unpaired Image Translation via Vector Symbolic Architectures

Theiss, Justin D.; Leverett, Jay; Kim, Daeil; Prakash, Aayush

doi:10.1007/978-3-031-19803-8_2

Cited by 20 publications

(8 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Baselines We compare with GAN-based image translation methods including MUNIT (Huang et al 2018), LPTN (Liang, Zeng, and Zhang 2021), CUT (Park et al 2020), TSIT (Jiang et al 2020), F-Sesim (Zheng, Cham, and Cai 2021) and VSAIT (Theiss et al 2022).…”

Section: Methodsmentioning

confidence: 99%

“…As shown in Figures 5-7, the MUNIT (Huang et al 2018), CUT (Park et al 2020), TSIT (Jiang et al 2020) and F-Sesim (Zheng, Cham, and Cai 2021) suffer from structural distortions and artifacts. Although LPTN (Liang, Zeng, and Zhang 2021) and VSAIT (Theiss et al 2022) can retain the structural features better (see Figures 5 and 6), they have limited style transfer ability. VSAIT (Theiss et al 2022) produces a satisfactory result when the source and target domains share similar styles (see Figure 7), whereas it still lacks style transfer ability.…”

Section: Qualitative Comparisonsmentioning

confidence: 99%

“…Although LPTN (Liang, Zeng, and Zhang 2021) and VSAIT (Theiss et al 2022) can retain the structural features better (see Figures 5 and 6), they have limited style transfer ability. VSAIT (Theiss et al 2022) produces a satisfactory result when the source and target domains share similar styles (see Figure 7), whereas it still lacks style transfer ability. In contrast, our method can faithfully preserve the structural details of the source image and fully capture the style of the target domain.…”

Section: Qualitative Comparisonsmentioning

confidence: 99%

“…In the work of LPTN (Liang, Zeng, and Zhang 2021) information. VSAIT (Theiss et al 2022) addresses semantic inversion issues by learning inverse mapping in a highdimensional vector space based on vector symbolic architectures to ensure consistency of source domain content. Although these methods have achieved success in maintaining image structure, the generated results still lack satisfactory style effects.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

SEIT: Structural Enhancement for Unsupervised Image Translation in Frequency Domain

Zhu,

Li,

et al. 2024

AAAI

View full text Add to dashboard Cite

For the task of unsupervised image translation, transforming the image style while preserving its original structure remains challenging. In this paper, we propose an unsupervised image translation method with structural enhancement in frequency domain named SEIT. Specifically, a frequency dynamic adaptive (FDA) module is designed for image style transformation that can well transfer the image style while maintaining its overall structure by decoupling the image content and style in frequency domain. Moreover, a wavelet-based structure enhancement (WSE) module is proposed to improve the intermediate translation results by matching the high-frequency information, thus enriching the structural details. Furthermore, a multi-scale network architecture is designed to extract the domain-specific information using image-independent encoders for both the source and target domains. The extensive experimental results well demonstrate the effectiveness of the proposed method.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Qualitative Comparisonsmentioning

confidence: 99%

Section: Qualitative Comparisonsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

SEIT: Structural Enhancement for Unsupervised Image Translation in Frequency Domain

Zhu,

Li,

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…Hyperdimensional Prototypes Refinement Recently, Hyperdimensional Computing has been used in computer vision tasks like few shot learning [17,26], outof-distribution detection [45], and image translation [42], which leverage quasi-orthogonal hyperdimensional representations without inducing much training and inference overhead. The initial hyperdimensional prototype (hp) is obtained based on the single-pass raw data as:…”

Section: Dual-prototype Self-augment and Refinementmentioning

confidence: 99%

Offline-Online Class-incremental Continual Learning via Dual-prototype Self-augment and Refinement

Huo¹,

Xu²,

Guo³

et al. 2023

Preprint

View full text Add to dashboard Cite

This paper investigates a new, practical, but challenging problem named Offline-Online Class-incremental Continual Learning (O 2 CL), which aims to preserve the discernibility of pre-trained (i.e., offline) base classes without buffering data examples, and efficiently learn novel classes continuously in a single-pass (i.e., online) data stream. The challenges of this task are mainly two-fold: 1) Both base and novel classes suffer from severe catastrophic forgetting as no previous samples are available for replay. 2) As the online data can only be observed once, there is no way to fully re-train the whole model, e.g., re-calibrate the decision boundaries via prototype alignment or feature distillation. In this paper, we propose a novel Dual-prototype Selfaugment and Refinement method (DSR) for O 2 CL problem, which consists of two strategies: 1) Dual class prototypes: Inner and hyper-dimensional prototypes are exploited to utilize the pre-trained information and obtain robust quasiorthogonal representations rather than example buffers for both privacy preservation and memory reduction. 2) Selfaugment and refinement: Instead of updating the whole network, we jointly optimize the extra projection module with the self-augment inner prototypes from base and novel classes, gradually refining the hyper-dimensional prototypes to obtain accurate decision boundaries for learned classes. Extensive experiments demonstrate the effectiveness and superiority of the proposed DSR in O 2 CL 1 .

show abstract

HVCLIP: High-Dimensional Vector in CLIP for Unsupervised Domain Adaptation

Vesdapunt,

Fu,

et al. 2024

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Unpaired Image Translation via Vector Symbolic Architectures

Cited by 20 publications

References 21 publications

SEIT: Structural Enhancement for Unsupervised Image Translation in Frequency Domain

SEIT: Structural Enhancement for Unsupervised Image Translation in Frequency Domain

Offline-Online Class-incremental Continual Learning via Dual-prototype Self-augment and Refinement

HVCLIP: High-Dimensional Vector in CLIP for Unsupervised Domain Adaptation

Contact Info

Product

Resources

About