Advanced Computer Vision with Deep-learning

Hello I am Nitsan Soffair, A Deep RL researcher at BGU.

In this Computer-vision course, you will learn the newest state-of-the-art Computer vision (CV) Deep-learning knowledge.

You will do the following

1. Get state-of-the-art knowledge of the following
   1. Object detection
   2. Image segmentation
   3. Visualization and Interpretability
2. Validate your knowledge by answering short and very easy 3-question queezes of each lecture
3. Be able to complete the course by ~2 hours.

Syllabus

1. Introduction to Computer vision
   1. Classification and Object detection

      Technology in the field of computer vision for finding and identifying objects in an image or video sequence

   2. Segmentation

      The process of partitioning a digital image into multiple image segments of pixels’ sets.

   3. Transfer-learning

      A research problem in machine learning (ML) that focuses on storing knowledge gained while solving one problem and applying it to a different but related problem.

   4. Resnets

      An artificial neural network (ANN). Skip connections are used to jump over some layers.

   5. Object localization

      a computer technology to detect instances of semantic objects of a certain class i.e. humans, buildings in images and videos.

2. Object detection
   1. R-CNN

      Detection algorithm.

   2. Fast R-CNN

      Detection network region-proposal algorithm.

   3. Faster R-CNN

      Object detection network region-proposal algorithm.

   4. RetinaNet

      A dense detector evaluating the loss.

3. Image segmentation
   1. FCN

      Transforms image pixels to classes using CNN.

   2. Upsampling methods

      Performed on a sequence of signal’s samples/continuous function.

   3. Evaluation with IoU and Dice-score

      Evaluation metrics.

   4. U-Net

      A Deep neural-networl model based on fully-connected neural-network.

4. Visualization and Interpretability
   1. Class activation maps

      Technique gets the discriminative image regions used by CNN to identify specific classes in image.

   2. Saliency maps

      An image that highlights the region on which people’s eyes focus first.

Resources

• Wikipedia
• Coursera