Tagged articles
image synthesis
27 articles · Page 1 of 1
The paper introduces TBSM, a lightweight direction‑tracking network that learns per‑sample Fake‑to‑Real vectors to guide a generator, enabling single‑forward (NFE=1) image synthesis with FID 1.92 on ImageNet‑512 and high‑quality text‑to‑image results, all without CFG, DMD, GAN, or drifting methods.
Generative AITBSMdiffusion models
0 likes · 11 min read

A²-Edit introduces a unified framework that handles arbitrary object categories and coarse masks through mixed‑Transformer expert routing, mask‑annealing training, and a 500K multi‑category dataset, achieving robust, high‑quality reference‑guided edits across e‑commerce, virtual try‑on, and visual effects scenarios.
UniEdit-500Kcross-category modelingexpert routing
0 likes · 10 min read

The drift model moves the distribution evolution into training to achieve single‑step image synthesis with near‑SOTA FID scores, but ICML reviewers gave it 5/4/4/3 and rejected it, citing heavy reliance on a pretrained feature encoder, unclear contribution of the drift field versus classifier‑free guidance, and insufficient comparison to prior work.
Classifier-Free GuidanceFIDICML rejection
0 likes · 9 min read

FlashAR transforms pretrained autoregressive image models into highly parallel generators, delivering up to 22.9× end-to-end speedup while using just 0.05% of the original training data and preserving generation quality, thanks to intermediate branching, a learnable fusion gate, and a two-stage adaptation process.
FlashAR+Model Accelerationautoregressive generation
0 likes · 10 min read

The rapid democratization of DeepFake technology lowers creation barriers, turning ordinary face photos into tools for malicious content, privacy breaches, and potential black‑market abuse, prompting urgent personal safeguards and regulatory action.
AI privacydeepfakeimage synthesis
0 likes · 13 min read

The article explains how JD.com's AI virtual try‑on system Oxygen Tryon uses advanced computer‑vision and generative models to let shoppers instantly preview clothing on their own photos, dramatically improving purchase decisions, reducing return rates, and outlining technical challenges, innovations, and future development plans.
AIComputer VisionFashion E‑commerce
0 likes · 7 min read

The article explains how JD.com's AI try‑on system Oxygen Tryon uses advanced computer‑vision models to let shoppers instantly preview garments on their own photos, dramatically improving fit perception, reducing return rates, and outlining future technical and business expansions.
AIComputer VisionFashion E‑commerce
0 likes · 6 min read

JD’s Oxygen Tryon leverages advanced AI, keypoint detection, and real‑time rendering to let shoppers virtually try on clothing, dramatically cutting return rates, boosting conversion, and outlining technical challenges, innovations, and future plans for broader fashion applications.
AI try-onComputer VisionFashion E‑commerce
0 likes · 6 min read

This article walks through the theory behind flow‑matching generative models, contrasting them with diffusion models, detailing the velocity‑field formulation, training objective, and sampling procedure, and includes visual illustrations of the core concepts.
Flow MatchingGenerative ModelsODE
0 likes · 8 min read


AIWalkerApr 2, 2025 · Artificial Intelligence
EasyControl introduces a lightweight condition‑injection LoRA module, a position‑aware training paradigm, and causal attention with KV‑cache to enable plug‑and‑play multi‑condition control for DiT models, supporting arbitrary image resolutions while cutting inference latency by up to 30% and preserving high‑quality generation.
Conditional GenerationDiTEasyControl
0 likes · 17 min read


AIWalkerMar 8, 2025 · Artificial Intelligence
IMAGPose introduces a unified conditional diffusion framework that combines feature‑level, image‑level, and cross‑view attention modules to generate high‑fidelity, photo‑realistic person images under diverse pose and multi‑view scenarios, outperforming prior SOTA methods on DeepFashion and Market‑1501.
AIComputer Visiondiffusion models
0 likes · 22 min read


AIWalkerJan 14, 2025 · Artificial Intelligence
The paper introduces DiC, a fully convolutional diffusion model that rethinks 3×3 convolutions, adds sparse skip connections, stage‑specific embeddings and conditional gating, and demonstrates superior FID/IS scores and faster inference compared to diffusion Transformers across multiple scales.
AIconvolutional networksdiffusion models
0 likes · 19 min read

In 2024, AI‑generated content transforms fresh‑food e‑commerce by leveraging large language and multimodal models to automatically craft concise product copy, synthesize realistic images and short‑video GIFs, detect and fix visual defects, and build brand‑specific style libraries, paving the way for future voice‑driven storytelling and personalized shopping experiences.
AIGCFresh E-commercebranding
0 likes · 7 min read


AntTechDec 19, 2024 · Artificial Intelligence
Framer is an interactive video frame interpolation method that leverages large‑pretrained video diffusion models, allowing users to define custom motion trajectories or use an automatic mode, and demonstrates strong performance in image deformation, video generation, and cartoon‑to‑video applications.
AIFramerdiffusion model
0 likes · 4 min read

SPLAM (Sub‑Path Linear Approximation Model) accelerates diffusion‑model image synthesis by linearly approximating short sub‑paths of the probability‑flow ODE, allowing high‑quality generation in as few as four steps, outperforming prior fast‑sampling methods on COCO benchmarks and being deployed in Alibaba Mama’s recommendation system.
AI image generationSPLAMdiffusion models
0 likes · 11 min read

This article presents JD Advertising's 2023 innovations that combine relation‑aware diffusion models, category‑aware background generation, and planning‑and‑rendering pipelines to automatically produce high‑quality, scalable, and personalized e‑commerce ad posters, addressing efficiency, cost, and creative limitations of manual design.
AIAdvertisingdiffusion
0 likes · 18 min read

The 360 AI Research Institute introduces HiCo, a hierarchical controllable diffusion model that enables fine‑grained layout control across up to eight image regions, integrates seamlessly with existing Stable Diffusion ecosystems, and demonstrates superior performance on the GRIT‑VAL benchmark for layout‑aware image synthesis.
AI drawingHiCocontrollable generation
0 likes · 8 min read

Since July, the author has been developing AIGC projects and observed dramatic image quality improvements driven by rapid advances in large models, open‑source plugins, and deeper understanding of generation pipelines and parameters, prompting a comprehensive summary of the Cyber project’s solutions, industry landscape, and front‑end responsibilities.
AI-generated imagesAIGCCyber project
0 likes · 8 min read

Google, in collaboration with UC Berkeley, has unveiled a novel generative AI approach called the Idempotent Generative Network (IGN) that can produce images from any input in a single step, offering an alternative to GANs, diffusion models, and consistency models.
GANGenerative AIIGN
0 likes · 3 min read

This guide introduces Tencent Hunyuan's upgraded text‑to‑image model, explains its technical innovations, provides detailed prompt engineering advice, showcases example prompts and generated images across various styles, and highlights real‑world applications and performance metrics for developers and creators.
AI GenerationTencent Hunyuanimage synthesis
0 likes · 12 min read

MiniGPT-5 is a novel multimodal generation model using generative vokens to interleave text and image synthesis, integrating Stable Diffusion and LLMs with a two-stage training that requires no domain-specific annotations, achieving state‑of‑the‑art coherence and quality on benchmarks like CC3M, VIST, and MMDialog.
AI researchMultimodal GenerationStable Diffusion
0 likes · 9 min read

The article demonstrates how to create visually appealing, scannable QR‑code artworks using ControlNet and Stable Diffusion, explaining QR‑code structure, contrast preservation, and several pipelines—including tile‑based, OpenPose‑combined, and community QR‑code models—while detailing WebUI settings, prompt examples, weight tuning, and a custom ControlNet that reduces grid artifacts.
AI-generated QR codesControlNetStable Diffusion
0 likes · 13 min read

Alibaba’s Creative & Video Platform showcases six newly accepted ACM MM and ECCV papers that introduce self‑supervised text‑erasing, a confidence‑driven action‑proposal module, a geometry‑aligned variational transformer for image‑conditioned layouts, a high‑resolution virtual‑try‑on system, a motion‑transformer for unsupervised animation, and a cross‑domain motion‑transfer framework, highlighting cutting‑edge AI for automated creative design, video editing, and e‑commerce applications.
Virtual Try-Onimage synthesismotion transformer
0 likes · 11 min read

Smart Creative at NetEase Yanxuan uses algorithmic templates to automatically generate personalized ad images and videos from product data, selecting colors, aesthetics, and sizes, reducing manual cost and boosting revenue across DSP, recommendation, and search advertising.
AI MarketingNetEase Yanxuanalgorithmic advertising
0 likes · 8 min read

This paper introduces a style‑adversarial autoencoder that separates content and style latent variables, uses a multi‑class discriminator, and demonstrates superior image generation and data augmentation across MNIST, face, and text datasets tasks.
Generative Adversarial Networksautoencodersdeep learning
0 likes · 15 min read

JD.com describes its AI-powered image automation pipeline—including intelligent cutout, layout learning, and batch synthesis—along with a large-scale product knowledge graph that enables applications such as the JIMI customer service robot and the Li Bai writing assistant for global e-commerce.
AICustomer Service RobotNatural Language Generation
0 likes · 7 min read

Alibaba’s AI Lab built an AR‑enhanced Starbucks coffee workshop in Shanghai, using client‑side object detection, deep‑learning cloud recognition, image synthesis, and color‑simulation techniques to overcome challenges like metal reflections, transparency, and varying lighting, illustrating how AR can revamp new‑retail experiences.
ARaugmented realitycloud recognition
0 likes · 8 min read
