Abstract Scene text detection methods based on neural networks have emerged recently and have shown promising results. Previous methods trained with rigid word-level bounding boxes exhibit limitations in representing the text region in an arbitrary shape. In this paper, we propose a new scene text detection method to effectively detect text area by exploring each character and affifinity between characters. To overcome the lack of individual character level annotations, our proposed framework exploits both the given characterlevel annotations for synthetic images and the estimated character-level ground-truths for real images acquired by the learned interim model. In order to estimate affifinity between characters, the network is trained with the newly proposed representation for affifinity. Extensive experiments on six benchmarks, including the TotalText and CTW-1500 datasets which contain highly curved texts in natural images, demonstrate that our character-level text detection signifificantly outperforms the state-of-the-art detectors. According to the results, our proposed method guarantees high flflexibility in detecting complicated scene text images, such as arbitrarily-oriented, curved, or deformed texts.