HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis

Face sketch-to-photo synthesis is crucial in law enforcement, converting forensic sketches into RGB images for criminal database matching. A major challenge is ensuring accurate color representation in synthesized images to avoid identification error caused by mismatched skin or eye color. However,...

Full description

Saved in:
Bibliographic Details
Main Authors: Yeji Choi, Haksub Kim, Kwanghoon Sohn, Ig-Jae Kim
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10916652/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850209860783702016
author Yeji Choi
Haksub Kim
Kwanghoon Sohn
Ig-Jae Kim
author_facet Yeji Choi
Haksub Kim
Kwanghoon Sohn
Ig-Jae Kim
author_sort Yeji Choi
collection DOAJ
description Face sketch-to-photo synthesis is crucial in law enforcement, converting forensic sketches into RGB images for criminal database matching. A major challenge is ensuring accurate color representation in synthesized images to avoid identification error caused by mismatched skin or eye color. However, direct sketch-to-photo translation struggles with proper color representation as it relies solely on grayscale sketches. While recent text-guided generative methods show promise for style adjustment based on text descriptions, they often produce mixed or exaggerated colors due to using a single representation for text prompts containing multiple entangled attributes. To address these challenges, we propose Hierarchical Text-guided Stylization (HiTS), a novel identity-preserving face sketch-to-photo synthesis method. HiTS categorizes text descriptions into intrinsic and mutable attributes, capturing both global and local color features. Using an encoder-decoder architecture, the encoder extracts global features from intrinsic attributes, while the decoder refines local styles via a semantic-textual embedding map. This map integrates text embeddings with facial parsing masks, enabling precise style adjustments for each facial component, even in small regions. Both quantitative and qualitative results demonstrate that HiTS achieves fine-grained stylization while preserving identity, leading to improved face recognition accuracy.
format Article
id doaj-art-0fe3a08bce0747a8bf18035b742b877e
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-0fe3a08bce0747a8bf18035b742b877e2025-08-20T02:09:55ZengIEEEIEEE Access2169-35362025-01-0113508855089410.1109/ACCESS.2025.354910210916652HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo SynthesisYeji Choi0https://orcid.org/0009-0001-8797-1576Haksub Kim1https://orcid.org/0000-0002-8780-9747Kwanghoon Sohn2https://orcid.org/0000-0002-3715-0331Ig-Jae Kim3https://orcid.org/0000-0002-2741-7047School of Electrical and Electronic Engineering, Yonsei University, Seoul, South KoreaArtificial Intelligence and Robotics Institute, Korea Institute of Science and Technology, Seoul, South KoreaSchool of Electrical and Electronic Engineering, Yonsei University, Seoul, South KoreaArtificial Intelligence and Robotics Institute, Korea Institute of Science and Technology, Seoul, South KoreaFace sketch-to-photo synthesis is crucial in law enforcement, converting forensic sketches into RGB images for criminal database matching. A major challenge is ensuring accurate color representation in synthesized images to avoid identification error caused by mismatched skin or eye color. However, direct sketch-to-photo translation struggles with proper color representation as it relies solely on grayscale sketches. While recent text-guided generative methods show promise for style adjustment based on text descriptions, they often produce mixed or exaggerated colors due to using a single representation for text prompts containing multiple entangled attributes. To address these challenges, we propose Hierarchical Text-guided Stylization (HiTS), a novel identity-preserving face sketch-to-photo synthesis method. HiTS categorizes text descriptions into intrinsic and mutable attributes, capturing both global and local color features. Using an encoder-decoder architecture, the encoder extracts global features from intrinsic attributes, while the decoder refines local styles via a semantic-textual embedding map. This map integrates text embeddings with facial parsing masks, enabling precise style adjustments for each facial component, even in small regions. Both quantitative and qualitative results demonstrate that HiTS achieves fine-grained stylization while preserving identity, leading to improved face recognition accuracy.https://ieeexplore.ieee.org/document/10916652/Face sketch-to-photo synthesisgenerative adversarial network (GAN)text-guided image generation
spellingShingle Yeji Choi
Haksub Kim
Kwanghoon Sohn
Ig-Jae Kim
HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis
IEEE Access
Face sketch-to-photo synthesis
generative adversarial network (GAN)
text-guided image generation
title HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis
title_full HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis
title_fullStr HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis
title_full_unstemmed HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis
title_short HiTS: Hierarchical Text-Guided Stylization for Face Sketch-to-Photo Synthesis
title_sort hits hierarchical text guided stylization for face sketch to photo synthesis
topic Face sketch-to-photo synthesis
generative adversarial network (GAN)
text-guided image generation
url https://ieeexplore.ieee.org/document/10916652/
work_keys_str_mv AT yejichoi hitshierarchicaltextguidedstylizationforfacesketchtophotosynthesis
AT haksubkim hitshierarchicaltextguidedstylizationforfacesketchtophotosynthesis
AT kwanghoonsohn hitshierarchicaltextguidedstylizationforfacesketchtophotosynthesis
AT igjaekim hitshierarchicaltextguidedstylizationforfacesketchtophotosynthesis