We present TNCR, a new table dataset with varying image quality collected from free open source websites. TNCR dataset can be used for table detection in scanned document images and their classification into 5 different classes.
2 PAPERS • NO BENCHMARKS YET