Dataset Groups Activity Stream CTW1500 dataset The CTW1500 dataset contains 1,500 images, of which 1,000 are for training and 500 are for testing. Each image has at least one curved text. BibTex: @dataset{Yuliang_et_al_2024, abstract = {The CTW1500 dataset contains 1,500 images, of which 1,000 are for training and 500 are for testing. Each image has at least one curved text.}, author = {Yuliang et al.}, doi = {10.57702/yve9p6ln}, institution = {No Organization}, keyword = {'arbitrary shapes', 'curved text', 'text detection'}, month = {dec}, publisher = {TIB}, title = {CTW1500 dataset}, url = {https://service.tib.eu/ldmservice/dataset/ctw1500-dataset}, year = {2024} }
