Convolutional Neural Networks

Explore a comprehensive lecture on deep learning for computer vision in this 35-minute video from MIT's Introduction to Deep Learning course. Dive into the fundamentals of convolutional neural networks, including their architecture, training process, and applications. Learn about evolutionary origins of vision, the visual cortex, image representation, and various computer vision tasks. Discover how CNNs extract features using convolution operations, pooling layers, and non-linearities. Examine advanced topics such as semantic segmentation, object detection, and image captioning. Gain insights into the impact of deep learning in face recognition, self-driving cars, and medicine. Understand the importance of data in computer vision applications and review key concepts covered in the lecture.

Intro
Vision: Evolutionary Origins
The Visual Cortex
Images are Numbers
Tasks in Computer Vision
High Level Feature Detection
Manual Feature Extraction
Learning Feature Representations
Fully Connected Neural Network
Using Spatial Structure
Applying Filters to Extract Features
Filters to Detect X Features
The Convolution Operation
Producing Feature Maps
Feature Extraction with Convolution
Convolutional Layers: Local Connectivity
Introducing Non-Linearity
Pooling
CNNs for Classification: Feature Learning
CNNs: Training with Backpropagation
ImageNet Dataset
ImageNet Challenge: Classification Task
An Architecture for Many Applications
Beyond Classification
Semantic Segmentation: FCNS
Driving Scene Segmentation
Object Detection with R-CNN
Image Captioning using RNNS
Class Activation Maps (CAM)
Data, Data, Data
Deep Learning for Computer Vision: Impact
Impact: Face Recognition
Impact: Self-Driving Cars
Impact: Medicine nature
Deep Learning for Computer Vision: Review