His graduate thesis doesn't feature a photo of a former playboy model

andrei?

suck my cock

ba tu!

(checked)
i'd like to convolute her kernel gaussian style

how many abortions did she have at the time of taking that photo?

Did you get it from Jow Forums?
I posted one I some months I converted from webp and scaled way down to get it < 3mb.

who cares? children are worthless and need to be dealt with one way or another.

OpenCV? Is there any demand for it?

can I have it, please?

I only use it when it has a function that numpy or scipy doesn't have, so almost never.

Is this some common picture that gets used for photo compression crap or something? We were given the same one a few weeks ago to use in my 4th year lab where we implemented some simple jpeg compression.

Attached: output.jpg (128x128, 8K)

yes, cs people are horny af

It's used in general throughout image processing, though it's been getting phased out in the Deep Learning community for some reason in place of pic related

Attached: baboon.jpg (1797x922, 589K)

It's something to do with the colors and tones that makes it useful as kind of a standard for calibrating shit

Hot.

So the equivalent of the teapot model for anything 3D rendering relate?

Yes, but it's stupid only using one picture as a reference. I'm almost convinced that ML papers use their training data as example pics in their research papers rather than validation/testing data.

Based and redpilled. I support fourth trimester abortion.

Must have scrolled through the folder 87 times.
dropbox.com/s/xucrmyg8vmthndq/15052202604.webp?dl=0
This shit's public domain, right?

Surely this rare beauty must have some value to computer sciences.
Funny how what was naughty enough for a playboy centerfold in 1956 won't even get you banned from a worksafe chinese funny pages board in current year.

Attached: 1956_12_Lisa Winters 01.jpg (1728x792, 154K)

Does my PC just not have enough power to tone map 60 fps 4k HDR in real time?

>2x Xeon X5650 2.67ghz
>GTX 980
>16gb ram

I tried it with madvr and got a good looking result, but the video frame rate is choppy. GTX 980 does not have hardware HEVC decoding, I think. If I converted the video into a different codec could I get better performance?

Anecdotally, I converted a 60 fps 4k HDR demo video to 2560x1600 (my monitor native resolution) 60 fps HDR HEVC and I was able to get stable, yet not completely smooth playback. But it took over 30 minutes to convert a 2 minute video.

Attached: good living.jpg (1024x768, 104K)

>thread about image processing
>reply with gay ass muh GPU shit

Jow Forums's a shit

whats the best machine learning model for image classification when dealing with a small dataset?

>to do with the colors and tones
horseshit, it was because she was hot, no other reason. that picture is not (and never was) very useful at all for 'tone comparison' - Lena was (and is) always pleasant to look at tho
somewhat conversly, you'd be fucking fired out of the average lab for using *that* as a demo/test object in about two seconds flat, in nowadays enlightened times, Pic Related

Attached: BurnTheWitch_.png (688x643, 620K)

>2x X5650
I remember running some benchmarks of my dual X5650 server against my ryzen 1700X, it got less than half the performance in shit like cinebench and video transcoding. They're 9 year old CPUs, you're just gonna have to deal with it.

find better ways to augment your data. Resnet or Inception are generally the gotos for image classification.

video processing is still image processing but this right here is true Jow Forums shit
doesn't answer the question just brags about their own hardware

As to the original question yes h.264 may play back better on a 980 but who knows how the hdr may be preserved in the conversion. Sounds like a rabbit hole not worth going down.

im using yolov3 atm which uses Darknet. What are your thoughts on these models?

>yolov3
meme-tier

whys that?

a

Here's the thread where they actually posted on Jow Forums
I found it on wasoru .org. You can also thank for sharing her name because honestly I didn't know it.

Also...
Why didn't jpeg2000 take off?

It's the meme of image processing classes.

That's not the right syntax.
>img
Guess he was right kek

I use it, but most CV is done with 3D sensors anyway, so PCL is much more useful.

Lmao it literally did. I made an extension that detected and flagged nude images using a neural network, using a dataset scraped off of /s/

Nice

Could someone recommend me a decent resource to get started with Image processing? Shit seems interesting.

I'm in physics, so very much proficient with Linalg and programming in Python, matlab, C++, java etc. I have no familiarity with image processing other than writing some basic motion tracking script for a ball falling on a high-contrast background.

jannies will delet this and probably ban me, but here you go

Attached: lenna_soderberg.jpg (1287x2800, 665K)

lol I also put it in 5 years ago
it was about image segmentation algorithms

For what it's worth that is the original for/from the magazine, but not the original reference image.
This one should be.
Janitors delete if you must but please don't enforce anything more.

Attached: Lena hi res.jpg (1084x2318, 557K)

from the archive not sure which came first

Attached: len_full.jpg (400x855, 158K)

I can't remember that pic being that hot tho