Hacker News new | past | comments | ask | show | jobs | submit login
OpenCV-Python Cheat Sheet: From Importing Images to Face Detection (fritz.ai)
177 points by salma-ghoneim 35 days ago | hide | past | web | favorite | 14 comments

PyImageSearch [1] is one of the best resources for doing actual stuff with opencv. I personally find the writing style a little patronising sometimes, but the content is excellent and kept up to date.

[1] https://www.pyimagesearch.com/

I'm not sure it's a cheat-sheet really, more a primer. Looks good though for those who've not used OpenCV (me!).

At this book you could find many good examples using opencv+python.


It explains concepts in a simple manner for beginners and it introduces cool stuff in little steps chapter by chapter.

(It is quite cheap, but probably you could also find the pdf version free on Internet)

I find this extremely unnerving to look at, let alone trying to process its contents. There is huge amounts of whitespace, huge fonts, huge images, single lines of text with more whitespace, a wide variety of font styles without any clear hierarchy. Hardly more than 3-4 sentences on the screen at a time. And of course useless, persistent header and footer banners that induce a feeling of looking through a closing garage door.

And then it hit me, this is just another TED talk style Medium publishing influencer.

Nothing of value here, just a bunch of random OpenCV examples you can just as well find in the official docs https://docs.opencv.org/master/d6/d00/tutorial_py_root.html

You cite a bunch of very valid criticisms, but then to say directly to the OP (as they posted their own article) that there is "nothing of value here" is excessive, IMHO, and symptomatic of a kind of discourse that only flies online (unless you would say that to a person's face when they presented you with something they created, and perhaps some people are that direct and lacking in empathetic communication).

Is there new information here? No. But it's presented much differently from the official documentation and perhaps some beginners would find a guide like this less intimidating, more consumable, whatever - people have different learning styles. Maybe the people that would prefer this tutorial over another are a small minority, but I take issue with the sweeping statement that the work the OP put in here created no value for anyone at all.

You are right, I would not say that to a person in meatspace and I should not say things like that to people on the internet either.


Looking at their other articles gives me the impression of a content creator that just pollutes the internet for keywords, potentially paid-for.

Just look at https://heartbeat.fritz.ai/ , which this random assortment of code snippets was compiled for, it's all just keyword stuffing as promotion for fritz.ai, plain and simple.

And oh look, nicely delayed submissions of article after article by OP: https://news.ycombinator.com/submitted?id=salma-ghoneim

Kudos for admitting a misstep and your follow-up analysis is definitely noteworthy. There was a suspicious ratio of upvotes to comments for this thread.

Why not post these observations instead, then? I think what you've written here is much more constructive than what was written above.

I'm someone who just took a computer vision course. I have other priorities (namely finals in other courses), so I haven't had much opportunity to repeatedly practice some of the more basic tasks in opencv. I wouldn't necessarily want to pour through my old assignment code to jump-start my memory on these tasks -- skimming through this now, I love how it's presented, and can see myself returning to this.

What you've linked targets a different use case. I don't need to relearn the core concepts through tutorials, and I don't need every possible config option from the docs. I just need simple, easy-to-parse demonstrations of basic use cases. This works for me.

>And then it hit me, this is just another TED talk style Medium publishing influencer.

"This" is a bit of an uncharitable way of framing that person, don't you think? Behind those blanket labels is an actual human with feelings and motivations that you and I don't actually know. Handwaving away a person's efforts like this is unnecessarily dismissive, imo. It smacks of "I don't like the motivations I've assumed from you, therefore your output doesn't matter." How would you feel it you were on the other end of that?

+1. I'm at the same noob stage and appreciate easy-to-digest courses. I, for one, am appreciative of the work.

FYI Cheatsheet is a concise set of notes used for quick reference. you opened an article called "OpenCV cheat sheet" expecting to find some novel published work? if you find it easy to read the documentation, go read it then.

Not sure if you’re having a different experience than I am. I’m reading this on a 9.7” iPad in a restaurant and it seems fine, so perhaps we have different first impressions.

I’ve written hundreds of tech articles and dozens of manuals, and I’ll tell you that people don’t want to read. Everyone’s busy. They don’t want a lot of theory or talk about the author’s life. They just want to get the job done. To me this article delivers beautifully on its promise.

^ This. Most folks these days need something short and to the point. If you want exhaustive detail, read the official documentation. If however you want something to point you in the right direction to get going with a couple of representative examples, then this style of presentation is great. If you subsequently need further detail, then you can search that out yourself. And let's be honest, searching for more is half the fun anyway ;-)

Nice job, this would have been nice to read through when i first started OpenCV.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact