Tell HN: App idea that should become reality | Hacker News Reader

Tell HN: App idea that should become reality

11 pointsstasy12 comments

I would pay for an app that allowed me to take a picture of a reading/pages of a book and get a summary for it. This would be incredibly useful for schoolwork. Basically like summly, but the summarized text isn't from some news site, but from the place you want it from (preferably a picture). If you make this and have it as an iPhone app, please tell me about it, and I will definitely buy it: aeip@live.com

View on HN

I would like an OCR app that, when launched, just OCR's the latest photo i've taken and put the resulting text in the clipboard.

Example use case: You're in a hotel and want to connect to the WiFi so instead of looking at their sign and typing in the password by hand, you just take a photo and launch the app and then paste the password where needed.

Would also be good for power users with apps like Launch Center and pythonista.

Or go one level down - I'd pay for a good OCR library and a good summarization library. The current OCR offerings are too difficult to get working well, and summarization tech is way too specialized at the moment.

What sort of challenges have you had with OCR? I'm working on an easy to use OCR sdk (iOS), we're going to start beta in a few weeks.

Tesseract is really difficult to tune such that you get decent results from camera images instead of cleanly scanned text.

Ocropus is really slow, but does seem to work better than tesseract out of the box. It's still quite bad with camera-captured images and IIRC doesn't give you any hooks to improve the results.

I couldn't find any other viable options when I was looking about six months ago.

I want an SDK that gives me an easy way to train the OCR engine. That means being able to view the images which performed badly, with text bounding boxes and the OCRed text overlaid that allow me to manually correct the results which it will then learn from. And, perhaps most importantly, good defaults for the common cases of document scans & camera captures.

This is an interesting idea, but what exactly do you mean by "summary"? Do you want an app to detect the most important parts of the text, and give you just those parts? Or are you interested in an app that detects what it is you would like summarized, and then find summaries on external sites like SparkNotes?

Not really a mobile developer myself, but Im guessing it would require parsing the text from a book with something like OpenCV, and then passing it through Summly's or clipped.me's API.

Where do you get Summly's API?

Even better - auto summary of everything you see through Google Glass.

I'd already be interested in one that just reads out the full text to me in a way that sounds natural enough that it isn't distracting.

With an ISBN it wouldn't be that hard to pull it from amazon.

With printings from school, you often get short passages or short stories that are not in books or very difficult to find within books. But just being able to take a picture of it would make it so much easier.

Identifying the text isn't that hard, generally 10+ word excerpts pasted into Google will give you a source. The reviews/summaries though... for books you have amazon and good reads, for short stories though I don't know where you would find them. The search: '"no advantages" review alice munro' for instance only turns up material related to the collection that it is a part of, not for the story.

btw, I'm kind of just brainstorming now.

Reply on news.ycombinator.com