# DeepFont: Identify Your Font from An Image

arXiv: https://arxiv.org/abs/1507.03196 · HTML: https://mlsanity.com/paper/1507.03196/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Jul 2015.

## Built on by (most cited first, 24 of 24)

- [De-rendering Stylized Texts](https://mlsanity.com/paper/2110.01890/) (2021)
- [Texture or Semantics? Vision-Language Models Get Lost in Font Recognition](https://mlsanity.com/paper/2503.23768/) (2025)
- [Impressions2Font: Generating Fonts by Specifying Impressions](https://mlsanity.com/paper/2103.10036/) (2021)
- [FontCLIP: A Semantic Typography Visual-Language Model for Multilingual Font Applications](https://mlsanity.com/paper/2403.06453/) (2024)
- [Font Generation with Missing Impression Labels](https://mlsanity.com/paper/2203.10348/) (2022)
- [CoSTAast: Cost-Sensitive Toolpath Agent for Multi-turn Image Editing](https://mlsanity.com/paper/2503.10613/) (2025)
- [FaSTA^*: Fast-Slow Toolpath Agent with Subroutine Mining for Efficient Multi-turn Image Editing](https://mlsanity.com/paper/2506.20911/) (2025)
- [Analyzing Font Style Usage and Contextual Factors in Real Images](https://mlsanity.com/paper/2306.12050/) (2023)
- [Font Style Interpolation with Diffusion Models](https://mlsanity.com/paper/2402.14311/) (2024)
- [HENet: Forcing a Network to Think More for Font Recognition](https://mlsanity.com/paper/2110.10872/) (2021)
- [CONSENT: Context Sensitive Transformer for Bold Words Classification](https://mlsanity.com/paper/2205.07683/) (2022)
- [From Fragment to One Piece: A Survey on AI-Driven Graphic Design](https://mlsanity.com/paper/2503.18641/) (2025)
- [TaCo: Textual Attribute Recognition via Contrastive Learning](https://mlsanity.com/paper/2208.10180/) (2022)
- [Parameter-Efficient Fine-Tuning of DINOv2 for Large-Scale Font Classification](https://mlsanity.com/paper/2602.13889/) (2026)
- [FONTNET: On-Device Font Understanding and Prediction Pipeline](https://mlsanity.com/paper/2103.16150/) (2021)
- [Learning Perceptual Manifold of Fonts](https://mlsanity.com/paper/2106.09198/) (2021)
- [Persis: A Persian Font Recognition Pipeline Using Convolutional Neural Networks](https://mlsanity.com/paper/2310.05255/) (2023)
- [Local Style Awareness of Font Images](https://mlsanity.com/paper/2310.06337/) (2023)
- [Total Disentanglement of Font Images into Style and Character Class Features](https://mlsanity.com/paper/2403.12784/) (2024)
- [How Could AI Support Design Education? A Study Across Fields Fuels Situating Analytics](https://mlsanity.com/paper/2404.17390/) (2024)
- [TexTAR : Textual Attribute Recognition in Multi-domain and Multi-lingual Document Images](https://mlsanity.com/paper/2509.13151/) (2025)
- [GRAPHIC--Guidelines for Reviewing Algorithmic Practices in Human-centred Design and Interaction for Creativity](https://mlsanity.com/paper/2511.17443/) (2025)
- [LogoDiffuser: Training-Free Multilingual Logo Generation and Stylization via Letter-Aware Attention Control](https://mlsanity.com/paper/2603.09759/) (2026)
- [FontFusion: Enhancing Generative Text in Diffusion Models with Typographic Conditioning](https://mlsanity.com/paper/2606.06066/) (2026)

## Abstract (arXiv)

As font is one of the core design concepts, automatic font identification and similar font suggestion from an image or photo has been on the wish list of many designers. We study the Visual Font Recognition (VFR) problem, and advance the state-of-the-art remarkably by developing the DeepFont system. First of all, we build up the first available large-scale VFR dataset, named AdobeVFR, consisting of both labeled synthetic data and partially labeled real-world data. Next, to combat the domain mismatch between available training and testing data, we introduce a Convolutional Neural Network (CNN) decomposition approach, using a domain adaptation technique based on a Stacked Convolutional Auto-Encoder (SCAE) that exploits a large corpus of unlabeled real-world text images combined with synthetic data preprocessed in a specific way. Moreover, we study a novel learning-based model compression approach, in order to reduce the DeepFont model size without sacrificing its performance. The DeepFont system achieves an accuracy of higher than 80% (top-5) on our collected dataset, and also produces a good font similarity measure for font selection and suggestion. We also achieve around 6 times compression of the model without any visible loss of recognition accuracy.