Dataset Groups Activity Stream Screen Point-and-Read (ScreenPR) benchmark The Screen Point-and-Read (ScreenPR) benchmark is a dataset for evaluating screen reading models. BibTex: @dataset{Yue_Fan_and_Lei_Ding_and_Ching-Chen_Kuo_and_Shan_Jiang_and_Yang_Zhao_and_Xinze_Guan_and_Jie_Yang_and_Yi_Zhang_and_Xin_Eric_Wang_2024, abstract = {The Screen Point-and-Read (ScreenPR) benchmark is a dataset for evaluating screen reading models.}, author = {Yue Fan and Lei Ding and Ching-Chen Kuo and Shan Jiang and Yang Zhao and Xinze Guan and Jie Yang and Yi Zhang and Xin Eric Wang}, doi = {10.57702/4os603r8}, institution = {No Organization}, keyword = {'GUI Understanding', 'Screen Reading'}, month = {dec}, publisher = {TIB}, title = {Screen Point-and-Read (ScreenPR) benchmark}, url = {https://service.tib.eu/ldmservice/dataset/screen-point-and-read--screenpr--benchmark}, year = {2024} }