You can adjust how many grayscale levels are distinguished (quantization). More levels mean a faster recognizer, although it may be less robust if the font is not a completely fixed target. Working on improving that too.
Either way, I think we have reached peak efficiency. It takes us exactly as long to recognize a letter in an image as it takes us to look up its string value in a tree map - both are O(log n). https://i.gazelle.rocks/beaHTML/37?quantize=2&attempts=10000 https://i.gazelle.rocks/beaHTML/37?quantize=2&attempts=10000 And what happens with https://i.gazelle.rocks/beaHTML/37?quantize=1 ? Yes, probably what you think will happen. Except the success message displayed is a bit too cocky... (thee is no notion of "unsolvable" yet). ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tef248b2dd124c952-Meb698d1fc5f0b1b56f8a2393 Delivery options: https://agi.topicbox.com/groups/agi/subscription
