diff options
| author | Soumith Chintala <soumith@gmail.com> | 2016-06-02 00:54:46 +0300 |
|---|---|---|
| committer | Soumith Chintala <soumith@gmail.com> | 2016-06-02 00:54:46 +0300 |
| commit | 572dc85f394fa5dfc2b92be7863ddb38883ec68f (patch) | |
| tree | 8750d98898d4b821b25da2ce90112402bf425378 | |
| parent | e553b7ae2698d44509031044d84ee3eac7c50141 (diff) | |
| parent | 0250b01dc61fbab97762f38f357bb3964416aa58 (diff) | |
Merge pull request #49 from szagoruyko/patch-1
Image links to heavy GIFs
| -rw-r--r-- | blog/_posts/2016-06-01-deep-fun-with-opencv.md | 18 |
1 files changed, 12 insertions, 6 deletions
diff --git a/blog/_posts/2016-06-01-deep-fun-with-opencv.md b/blog/_posts/2016-06-01-deep-fun-with-opencv.md index 26976e9..37b4cdd 100644 --- a/blog/_posts/2016-06-01-deep-fun-with-opencv.md +++ b/blog/_posts/2016-06-01-deep-fun-with-opencv.md @@ -20,7 +20,7 @@ Usage Examples A basic example may be live CNN-based image classification. In the following demo, we grab a frame from the webcam, then take a central crop from it and use a small ImageNet classification pretrained network to predict what's in the picture. Afterwards, the image itself and the 5 most probable class names are displayed. - +[](http://cdn.makeagif.com/media/2-28-2016/p4xoRF.gif) The comments should explain the code well. *Note: this sample assumes you already have the trained CNN; see the [original code on GitHub](https://github.com/szagoruyko/torch-opencv-demos/blob/master/imagenet_classification/demo.lua) by Sergey Zagoruyko that automatically downloads it.* @@ -112,15 +112,21 @@ Of course, this is quite an inefficient way of face detection provided just to b The entire code is [available on GitHub](https://github.com/szagoruyko/torch-opencv-demos/blob/master/age_gender/demo.lua). Here's a sample from IMAGINE Lab by Sergey Zagoruyko: - + -[And here is is a heavy just-for-fun GIF](https://raw.githubusercontent.com/torch/torch.github.io/master/blog/_posts/images/opencv_age.gif). +And here is a heavy just-for-fun GIF: + +[](http://cdn.makeagif.com/media/3-11-2016/afVDJO.gif) ### NeuralTalk2 A good image captioning example is [NeuralTalk2](https://github.com/karpathy/neuraltalk2) by Andrej Karpathy. With OpenCV, it's easy to make this model caption live video or camera stream: - +[](http://cdn.makeagif.com/media/4-04-2016/eLgBBZ.gif) + +[](http://cdn.makeagif.com/media/4-04-2016/92PJ5o.gif) + +[](http://cdn.makeagif.com/media/4-04-2016/7ysYNO.gif) Here are a couple more demos: [one](https://raw.githubusercontent.com/torch/torch.github.io/master/blog/_posts/images/opencv_neuraltalk_demo2.gif) and [another](https://raw.githubusercontent.com/torch/torch.github.io/master/blog/_posts/images/opencv_neuraltalk_demo3.gif). @@ -218,9 +224,9 @@ The whole runnable script is [available here](https://github.com/shrubb/torch-op The [Texture Networks: Feed-forward Synthesis of Textures and Stylized Images](http://arxiv.org/abs/1603.03417) paper proposes an architecture to stylize images straightforwardly, shipping with an [open source implementation in Torch](https://github.com/DmitryUlyanov/texture_nets/). It takes ~20 ms to process a single image with Tesla K40 GPU, and ~1000 ms with CPU. Having this, a tiny modification allows us to render any scene in a particular style in real time: - +[](http://i.makeagif.com/media/4-24-2016/0zb-UY.gif) - +[](http://i.makeagif.com/media/4-23-2016/Pk4ZAL.gif) By the way, these very GIFs (originally in form of encoded videos) were rendered using OpenCV as well. There is a `VideoWriter` class that serves as a simple interface to video codecs. Here is a sketch of a program that encodes similar sequence of frames as a video file and saves it to disk: |