Crowdsourced Report Generation via Bug Screenshot Understanding

Quality control is a challenge of crowdsourcing, especially in software testing. As some unprofessional workers involved, low-quality yieldings may hinder crowdsourced testing from satisfying requesters' requirements. Therefore, it is in demand to assist crowdworkers to raise bug report quality. In this paper, we propose a novel auxiliary method, namely CroReG, to generate crowdsourcing bug reports by analyzing bug screenshots uploaded by crowdworkers with image understanding techniques. The preliminary experiment results show that CroReG can effectively generate bug reports containing accurate screenshot captions and providing positive guidance for crowdworkers.