5 Super Cool Computer Vision Applications Using Deep Learning

In today’s video I’m going to be talking about 5 really cool computer vision applications that can be implemented using deep learning.

Computer vision is an area in computer science which focuses on enabling computers to see, identify and process images in the same way that humans do.

Deep learning is a machine learning technique that teaches computers to do what comes naturally to us humans: learning by example. In deep learning, a computer mode is trained by using a large set of labeled data and neural network architectures that contain many layers.

5. Image Classification:

Image classification is the MOST popular use of computer vision. It is the process of identifying objects in images. For Human beings when we see a picture, we know automatically what it is. Its not that simple for a computer. And this problem gets harder to solve when we are working with huge data. Image classification is used in the auto industry(in self driving cars) and healthcare(for instance identifying tumors from endoscopies & biopsies).
4. Image Colorization:

Image Colorization does exactly what it says; it adds color to B&W images. However achieving the real color of an image is extremely difficult, and here is why. When we have a grey scale image, we essentially only have one kind of information of each of the pixels; their intensity. So how do we determine the RGB value of each of the pixels? Turns out each pixels has 313 different classes(Colours). So the problem with image classification is that each grey pixel has 313 color possibilities to choose from.

3. 3D Reconstruction:

For a human, it is usually an easy task to get an idea of the 3D structure shown in an image. Due to the loss of one dimension in the projection process, the estimation of the true 3D geometry is difficult and a so called ill-posed problem, because usually infinitely many different 3D surfaces may produce the same set of images.

2. Image Synthesis:

Image synthesis is an extremely broad genre. Most forms of image synthesis use what’s called a Generative Adversial Network(AKA GAN) and it is a type of neural network for generating models. GANs generate images that are similar to an existing image but not exactly the same. For instance, it can be used to generate human faces.

1. Image Style Transfer:

Is a technique using 2 images(“content” & “style reference”) and blending them together so the output image looks like the content image but painted on like the style reference image. It can be used in Art, In computer graphics, media and many more.
Music: bensound.com

Видео 5 Super Cool Computer Vision Applications Using Deep Learning канала Smitha Kolan - Machine Learning Engineer