태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
Font recognition (also called visual font recognition or optical font recognition) is the task of identifying the font family or families used in images containing text. Understanding which fonts are used in text may, for example, help designers find the right style, as well as help select an optical character recognition engine or model that is a better fit for certain texts.
출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.
Multi-Modal Reading (MMR) Benchmark includes 550 annotated question-answer pairs across 11 distinct tasks involving texts, fonts, visual elements, bounding boxes, spatial relations, and grounding, with carefully designed evaluation metrics.
Explor_all font image dataset https://drive.google.com/file/d/1P2DbNbVw4Q__WcV1YdzE7zsDKilmd3pO/view
Subset of AdobeVFR. The dataset contains "real-world text images". > We collected 201,780 text images from various typography forums, where people post these images seeking help from experts to identify the fonts. Most of them come with hand-annotated font labels which may be inaccurate. (...) Finally, we obtain 4,384 real-world test images with reliable labels, covering 617 classes (out of 2,383). (...) Removing the 4,384 labeled images from the full set, we are left with 197,396 unlabeled real- world images which we denote as VFR_real_u. The labeled images form VFR_real_test. The other part of AdobeVFR consists of synthetic data (with 2383 classes).
Subset of AdobeVFR. The dataset contains images depicting English text and consists of 1000 synthetic images for training and 100 for testing, for each of 2383 font classes. The training and test sets are called VFR_syn_train and VFR_syn_val, respectively. The other part of AdobeVFR consists of "real-world text images".
325 word images intended for font recognition, whose fonts are included in [VFR-447] (and [VFR-2420]). > (...) 325 real world test images for the font classes we have in the training set. These images were collected from typography forums, such as myfonts.com, where people post these images seeking help from experts to identify the fonts. Compared with the synthetic data, these images typically have much larger appearance variations caused by scale, background, lighting, noise, perspective distortions, and compression artifacts. We manually cropped the texts from these images with a bounding box to normalize the text size approximately to the same scale as the synthetic data. [VFR-447]: https://paperswithcode.com/dataset/vfr-447 [VFR-2420]: https://paperswithcode.com/dataset/vfr-2420
Persian Font Recognition (PFR) A dataset in order to solve font recognition for the Persian language. This dataset is part of a paper titled "Persis: A Persian Font Recognition Pipeline Using Convolutional Neural Networks". If you use the findings of the paper or this dataset please cite the paper. GitHub repo
Persian Text Image Segmentation (PTI SEG) This dataset is part of a paper titled "Persis: A Persian Font Recognition Pipeline Using Convolutional Neural Networks". A dataset in order to solve image segmentation for the Persian texts. For generating the PTI SEG dataset, we use 735 different backgrounds in four types (stock, papers, noisy real world, and texture), and apply four effects (gradient light, folding, subtle noise, and ink bleed) to the image If you use the findings of the paper or this dataset please cite the paper. GitHub repo
A synthetic dataset containing word images of 447 typefaces with font variations for each typeface, created for visual font recognition. > We collect in total 447 typefaces, each with different number of variations resulting from combinations of different styles, e.g., regular, semibold, bold, black, and italic, leading to 2,420 font classes in the end. > Each class in VFR-447 and VFR-2420 has 1,000 synthetic word images, which are evenly split into 500 training and 500 testing. There are no common words between the training and testing images. > To model the realistic use cases, we add moderate distortions and noise to the synthetic data.
A synthetic dataset containing 447 typefaces with only one font variation for each typeface, created for visual font recognition. > Each class in VFR-447 and VFR-2420 has 1,000 synthetic word images, which are evenly split into 500 training and 500 testing. There are no common words between the training and testing images. > To model the realistic use cases, we add moderate distortions and noise to the synthetic data.
This is a dataset of 18624 fonts labeled as 100% Free and Public domain / GPL / OFL on https://www.dafont.com/ with .ttf and .otf extensions. Code used to create it can be found at: https://github.com/duskvirkus/dafonts-free
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.