Offline handwritten mathematical expression recognition is often considered much harder than its online counterpart due to the absence of temporal information and the presence of background noise. In order to take advantage of the more developed techniques on online recognition and save resources, an oversegmentation approach is proposed to recover strokes from a textual bitmap image automatically. The proposed algorithm first break down the skeleton of a binarized image into junctions and segments, then segments are merged to form strokes, finally the ordering is determined by recursive projection and topological sort. Given a state-of-art online handwritten mathematical expression recognition system, the proposed procedure correctly recognized 58.22%, 65.65% and 65.05% of the offline formulas rendered from CROHME 2014, 2016 and 2019 respectively. Therefore, the effectiveness of stroke extraction to offline recognition is justified.