BlindVision

Getting frustrated by the low accuracy levels of good ‘ol Raw Code Processing or even the (way) more advanced Computer Vision, HIPSTO’s data science team tasked themselves to come up with something faster, more accurate and cost efficient. A triple dare with a whopping outcome. After many months of deep-tech R&D, HIPSTO is about to market its proprietary AI web data feature extraction and labeling/annotation technology coined BlindVision.

To put BlindVision to the test, we engaged in a showdown against Computer Vision (considered the most advanced solution to date). Both extraction-annotation pipelines were inputted with the same 1200 (randomly chosen) web article URLs in order to extract and label 5 specific elements. BlindVision came out on top -time and time again- as we ran the test multiple times. Test results showed a +35% in accuracy at -65% in cost for continuous monitoring. Yiihaaaa!!

We’re now testing the waters for 2 specific use cases (Alt Data for financial markets and Brand Safety for the media & publishing industry),……Much more to come!

Drop us a line to get a live demo or discuss a use case!