Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 7 years ago.
Improve this question
Sorry if this has been addressed before, I searched it up and I couldn't seem to find a question like this. I am making a social media app, and I want to preserve the quality of the images as much as possible. To make a data representation of the image I am using "UIImageJPEGRepresentation" and I would like to know what an optimal setting for the compression quality is.
One big consequence of setting the quality below 1 is that the image uploads to my server in much less time. I have experimented with the compression and to me I can't really tell the difference between 0.6 and 1 unless I zoom in on a computer, but I just wanted to know if there was a number or range that would produce favorable results.
One of the interesting and evolving format is Webp format that Google has introduced. This article suggests that Facebook is also trying to use this format. To answer the exact question as to how much to compress please note the following:
-Format of the image(I assume yours is JPEG)
-Compression technique(Lossy or Lossless)
-Target devices(I assume yours is mobile)
Considering above parameters(and more) and looking at the dimensions of all social networking sites I suggest you to get the compression level high enough such that you can see a difference in the image quality on the computer, that way you would have found an optimal level, remember lower the better till it degrades. Additionally you can find the information at this.
Related
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 4 years ago.
Improve this question
Should webgl be used for simple websites?
I'm not sure or it is wise to use webgl for a simple website just to give it a better look. Will this work on all devices?
WebGL is widely supported today https://www.caniuse.com/#feat=webgl
Whether you "should" use it or not is a broad question. Remember that you aim at improving the user experience. People are forgiving when they play video games, but they don't want to hear their computer fans spin, witness their battery discharging very fast or feel their device getting hot when all they wanted was to read a cooking recipe. Try to be user friendly.
You may for instance want to cap the framerate and/or reduce the resolution on high definition devices, pause the animation loop when the window looses focus (which is not the default behaviour of requestAnimationFrame) or when there is nothing changing on the screen (if the WebGL element is interactive for example). Also, try to write efficient algorithms: it's easy to start writing things on the fragment shader or the CPU when they should be done on the vertex shader. There are many ways to accomplish the same thing and they don't put the same stress on the computer.
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 5 years ago.
Improve this question
I am hoping that TensorFlow can turn this input, to this output.
Input: A floorplan PNG, and 1 - 5 images of a symbol
Output: The same floorplan, but with all matching symbols highlighted
I can do the hard work of figuring out HOW to do it, but I don't want to waste 2 weeks just to figure out it wouldn't be possible. I know I'd need to train it with multiple images, but I won't have more than 5 examples of a given symbol.
Does TensorFlow have these capabilities?
Thanks!
Yes, it is possible to use tensorflow to create a machine learning algorithm to do that for you, but I would bet that is not how you want to do this. First off, in order to do this in tensorflow, you would need to manually create a large number of training samples and spend a significant amount of time figuring out how to define the network and train it. Sure, you could do it, but I definitely wouldn't advise it.
If you have a specific set of symbols that you want to highlight, it would probably be better to use opencv to find and highlight the symbols. For example, in opencv, you could use Template Matching to find a specific symbol in the floor plan and then highlight them by modifying pixel color.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 9 years ago.
Improve this question
I am creating an iPad app for golfers, I will get their score card image as below. I want to calculate sum of scores for each person by scanning this image.
Is there any way. Please give me a logic for this.
I have used opencv http://opencv.org in the past to do something similar but with sudoku puzzles.
It is a LOT of work and making it work with handwriting will add to the difficulty.
I found a really good resource for analysing sudoku grids. I'll try to find it again but it was 4 years ago.
Good luck though.
There is a Tesseract port for iOS which is about the best OCR you're likely to get on the platform without either:
A) Porting another OCR library
or
B) Shipping the images off to an online OCR service
To make this more complex, you don't just want to OCR but you want to OCR handwriting and put it into a grid. This is not something that can be done overnight but is in fact rather complex. Not impossible, but complex.
Would it not just be easier to let the players enter there scores straight onto the app and then airprint a score card?
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 1 year ago.
Improve this question
Are there any libraries, in any language, out there to help identify and grab the images of people in a still photo? Something similar in effect to the way the Kinect can isolate users.
Thanks much!
I think it depends very much on the setup (e.g. simple bg. with decent lighting condition vs. random bg. with random lighting). If you can make life easier for yourself and isolate a few simpler use cases that would be great. Still there are other available method, look at the plethora of research around pedestrian detection for example.
One thing I did try and it works surprisingly well although computationally intensive is the Histogram of Gradient Orientations, implemented in OpenCV as the HoG descriptor. For a still photo this should produce decent results. You can have a look at the OpenCV sample. I also recommend having a look at Dramanan's excellent papers.
Long story short, thanks for years of inspiring research in computer vision, there are quite a few interesting options out there, it's up to how willing you are to go into detail. Still, regardless of how clever algorithms can be, I believe it's far more important to get a decent setup that allows simple and efficient solutions rather than complex solutions that try to cater for every possible situation. Goodluck!
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 7 years ago.
Improve this question
Given n number of raw URLs, I'd like to be able to classify them by:
news, blog, photo and video.
An example would be if a link directs a user to a photo, would it be enough to say that the raw link contains file extension for images to be able to classify the raw URL as photo?
As for video, blog and news, it seems it isn't enough to have a set of domains (like http://www.youtube.com) that will classify the raw URLs.
Could classification be done by examining the web content? Or are there any open source tools for this?
The only URLs that may be even somewhat reliably classified, are those that point to a distinct medium (i.e. http://foo.com/foo.jpg is most certainly an image). Otherwise, you must analyze the content of the page.
This can be a bit tricky, as Flash may contain a photo, video, or neither, without providing any searchable clue as to the content of the flash object. With enough effort, this can obviously be overcome (Google does it!), but I'm not aware of any open source resources that provide a library of media-related domains. Such data result from countless programmer-hours of effort -- an effort that typically seeks a return on investment (ROI). Case in point, ClueWeb09 is just a dataset of downloaded pages, used to test search algorithms -- not really sorted or categorized.
"Sometimes no help is the answer."