DeepFont: Identify Your Font from An Image

By Zhangyang Wang, Jianchao Yang, Hailin Jin, Eli Shechtman, Aseem Agarwala, Jonathan Brandt, Thomas S. Huang. 2015

136 citations recorded by the source corpus.

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 [4] LFE, 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.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Studying Very Low Resolution Recognition Using Deep Networks (2016, methodology): “That outperforms the best accuracy numbers reported on VFR before, 38.15%@top-1 and 20.62%@top-5 [33].”
DADA: Deep Adversarial Data Augmentation for Extremely Low Data Regime Classification (2018, background): “A number of works [28–33] explored the “free” generation of labeled synthetics examples to assist training.”
Image Captioning with Deep Bidirectional LSTMs (2016, background): “Additionally, it has been proved that data augmentation such as randomly cropping and horizontal mirror [32, 22], adding noise, blur and rotation [40] can effectively alleviate over-fitting and other.”
Benchmarking Single-Image Dehazing and Beyond (2017, background): “For example, [61] provided an example of pre-training the robust low-level CNN filters using”
Controllable Artistic Text Style Transfer via Shape-Matching GAN (2019, background): “2(b) for example, one may see that the glyph with more complex strokes is more vulnerable to large glyph deformation [28].”
EnergyNet: Energy-Efficient Dynamic Inference (2018, background): “Most of them aim to reduce the model size or the number of 28 computations [3, 4, 5, 6, 7, 8, 9, 10, 11, 12].”
Deep k-Means: Re-Training and Parameter Sharing with Harder Cluster Assignments for Compressing Deep Convolutions (2018, methodology): “…CNN compression techniques have been proposed, such as weight compression (Bhattacharya & Lane, 2016) (Han et al., 2016) (Lane et al., 2016) and decomposition (Changpinyo et al., 2017) (Howard et al., 2017) (Wang et al., 2015), and compact architectures (Iandola et al., 2016) (Lin et al., 2014).”