Fréchet inception distance
PulseAugur coverage of Fréchet inception distance — every cluster mentioning Fréchet inception distance across labs, papers, and developer communities, ranked by signal.
- used by Diffusion Transformer 70%
- used by Structural Similarity Index Measure 70%
- used by Diffusion Models 70%
- used by ControlNet 70%
- used by Latent diffusion model 70%
- competes with Kernel Inception Distance 70%
- used by lpips 70%
- instance of Structural Similarity Index Measure 60%
- used by DagsHub 60%
- competes with lpips 60%
- instance of U-Net 60%
- instance of alphaXiv 60%
4 day(s) with sentiment data
-
PDA++ framework enhances remote sensing object insertion for few-shot learning
Researchers have developed PDA++, a novel framework for realistic object insertion in remote sensing imagery. This system aims to enhance few-shot learning and address data scarcity by generating synthetic targets that …
-
AI generates culturally faithful Ulos motifs using multimodal Stable Diffusion XL
Researchers have developed a multimodal generative framework to aid the traditional Batak Ulos weaving industry by enabling controllable and culturally faithful motif generation. The system fine-tunes Stable Diffusion X…
-
New SANI framework enhances diffusion models with pixel-level noise injection
Researchers have introduced Spatially Adaptive Noise Injection (SANI), a new framework for diffusion models that dynamically adjusts noise application on a per-pixel basis. Unlike traditional methods that apply noise un…
-
New AI method enhances image outpainting for subject clarity
Researchers have developed a new framework for image outpainting that aims to improve the clarity and fidelity of the primary subject within an image. This method combines vision-language model (VLM) guidance with multi…
-
New research tackles diffusion model safety and data generation
Researchers are developing advanced techniques to improve the safety and utility of diffusion models. One approach, ".ours", focuses on generating synthetic tabular data that not only matches real data distributions but…
-
New VTON evaluation framework DAT outperforms Gemini, Qwen, and GPT-5.5
Researchers have developed a new framework called DAT to evaluate virtual try-on (VTON) models more effectively. Existing metrics like FID and SSIM struggle to capture garment fidelity, so DAT breaks down consistency in…
-
StyleGANCA: Lightweight NCA for Medical Image Synthesis
Researchers have introduced StyleGANCA, a novel generative adversarial network that utilizes Neural Cellular Automata (NCA) for medical image synthesis. This architecture integrates a StyleGAN-inspired mapping network w…
-
Mixture-Greedy strategy outperforms UCB for generative model selection
A new research paper proposes a simpler 'Mixture-Greedy' strategy for selecting generative models, challenging the necessity of Upper Confidence Bound (UCB) bonuses in diversity-aware multi-armed bandit tasks. The study…
-
New method uses metadata to guide synthetic cardiac MRI generation
Researchers have developed a new method for generating synthetic cardiac magnetic resonance imaging (CMR) using a pre-trained latent diffusion model. This approach conditions the model on structured clinical metadata an…
-
New ZID metric offers improved evaluation for generative models
Researchers have introduced ZID (Z-resolved Integrated Diagnostic), a new evaluation metric for generative models that aims to improve upon existing metrics like Fréchet Inception Distance (FID) and Kernel Inception Dis…
-
Ordinal diffusion model generates realistic medical images with ordered disease progression
Researchers have developed an ordinal latent diffusion model designed to generate color fundus images, specifically addressing the continuous nature of disease progression in ophthalmology. Unlike standard conditional d…
-
New AI method restores postmortem tissue images for forensic diagnostics
Researchers have developed a new method to restore degraded images of postmortem tissue samples, aiming to improve forensic diagnostics. This technique addresses the challenge of image degradation caused by autolysis, a…
-
New method calibrates generative model training paths for improved performance
Researchers have introduced Difficulty-Calibrated Flow Matching, a novel approach to training generative models. This method dynamically adjusts the noise-to-data interpolation path based on the model's learning difficu…
-
New OccluRank framework enhances image generation with ordinal occlusion control
Researchers have introduced OccluRank, a novel framework for layout-to-image generation that enhances control over occlusion by incorporating a simple ordinal rank for each bounding box. This method allows users to spec…
-
New MeanFlow-Transfer method accelerates generative model training
Researchers have developed a new method called MeanFlow-Transfer (MF-T) to accelerate the training of generative models on new domains with limited data. This approach unifies adaptation and acceleration by mapping dive…
-
New diffusion model generates controllable, high-risk driving scenarios
Researchers have developed RiskMV-DPO, a novel pipeline for generating safety-critical driving scenarios to enhance autonomous driving systems. This method allows for risk-controllable multi-view scenario generation by …
-
New research offers deployment-aware order for CycleGAN enhancements
A new research paper proposes a deployment-aware adoption order for enhancements to Cycle-Consistent Adversarial Networks (CycleGANs), a type of generative model used for image-to-image translation. The paper identifies…
-
New research questions Fréchet Inception Distance trustworthiness
A new paper published on arXiv explores the trustworthiness of the Fréchet Inception Distance (FID) metric, commonly used to evaluate synthetic image quality. The research, authored by Ciaran Bench, investigates how sto…
-
Friction-Augmented Drifting Models enhance resource-efficient domain translation
Researchers have introduced Friction-Augmented Drifting Models (DMF), a novel approach to domain translation that significantly enhances resource efficiency. DMF addresses limitations in existing Drifting Models (DMs) b…
-
New LOB-ID framework evaluates synthetic market data using AI embeddings
Researchers have introduced LOB-ID, a novel framework for evaluating synthetic market data generated by AI models. This framework adapts existing embedding-based metrics like Fréchet Inception Distance (FID) and Monge I…