Classification in Computer Vision - Part II - Lecture 19

Explore advanced classification techniques in computer vision through this comprehensive lecture from the University of Central Florida's CAP5415 Computer Vision course. Delve into support vector machines (SVMs), including nonlinear SVMs and their advantages. Examine the machine learning framework, feature extraction, and neural networks, with a focus on fully convolutional networks and the conversion of fully connected layers to convolutional layers. Learn about various activation functions, binary and multi-label classification, loss functions, and optimization techniques like gradient descent. Gain insights into network training processes and visualize convolutional operations. This in-depth lecture equips you with essential knowledge for tackling complex classification problems in computer vision applications.

Intro
SV Classifier
Nonlinear SVMS
SVMS: Pros and cons
The machine learning framework
Features
Neural Networks
Fully convolutional network
Converting FC into conv
Activation Functions
Binary classification
Softmax activation
Multi-label
Loss Function
Train with Gradient Descent
Optimization
Network training
UCF Visualizing Convolution