This paper introduces several extractive approaches for automatic image tagging, relying exclusively on information mined from texts. Through evaluations on two datasets, we show that our methods exceed competitive baselines by a large margin, and compare favorably with the stateof-the-art that uses both textual and image features.