FLUXSynID: A Synthetic Face Dataset with Document and Live Images
Abstract
FLUXSynID: A Synthetic Face Dataset with Document and Live Images FLUXSynID is a high-resolution synthetic identity dataset containing 14,889 unique synthetic identities, each represented through a document-style image and three live capture variants. Identities are generated using the FLUX.1 [dev] diffusion model, guided by user-defined identity attributes such as gender, age, region of origin, and other various identity features. The dataset is created to support biometric research, including face recognition and morphing attack detection. File Structure Each identity has a dedicated folder (named as a 12-digit hex string, e.g., 000e23cdce23) containing the following 5 files: 000e23cdce23_f.json — metadata including sampled identity attributes, prompt, generation seed, etc. (_f = female; _m = male; _nb = non-binary) 000e23cdce23_f_doc.png — document-style frontal image 000e23cdce23_f_live_0_e_d1.jpg — live image generated with LivePortrait (_e = expression and pose) 000e23cdce23_f_live_0_a_d1.jpg — live image via Arc2Face (_a = arc2face) 000e23cdce23_f_live_0_p_d1.jpg — live image via PuLID (_p = pulid) All document and LivePortrait/PuLID images are 1024×1024. Arc2Face images are 512×512 due to original model constraints. Attribute Sampling and Prompting The attributes/ directory contains all information about how identity attributes were sampled: A set of .txt files (e.g., ages.txt, eye_shape.txt, body_type.txt) — each lists the possible values for one attribute class,…
Read the paper
Related studies
- Learning from First-Generation Qualitative Approaches in the IS Discipline: An Evolutionary View and Some Implications for Authors and Evaluators (PART 1/2)
- Commentary on the clinical and preclinical dosage limits of interstitially administered magnetic fluids for therapeutic hyperthermia based on current practice and efficacy models
- Prosthesis: New Technological Opportunities and Innovative Biomedical Devices
- The history and future of neural modeling for cochlear implants
- Ryanodine receptors: waking up from refractoriness
- Novel Transgenic Rabbit Model Sheds Light on the Puzzling Role of Matrix Metalloproteinase-12 in Atherosclerosis
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.