Explore video content understanding using text in this 44-minute lecture by Amir Mazaheri from the University of Central Florida. Delve into the challenges, algorithms, and training methods for video representation and scoring functions. Learn about exemplar queries, sentence encoding, spatial attention networks, and temporal attention models. Examine inaccuracy detection and correction techniques, as well as the opposite problem of generating video from text. Discover the proposed approach using generator and discriminator block diagrams, conditional batch normalization, and frame generation. Analyze evaluation metrics, quantitative results, and potential future work in this comprehensive overview of video content analysis and generation techniques.
Overview
Syllabus
Intro
Motivation
Challenges
Algorithm
Training
Video Representation
Scoring function
Optimization - Updating Rules
Exemplar queries
Test on Unseen Queries
Qualitative results
Sentence Encoder
Spatial Attention Network • Which regions of the frames to look?
Temporal Attention Model
Inference Module
Experiments
Limitations
What is an Inaccuracy?
Formulation
Detection By Reconstruction
Visual Features
Inaccuracy Detection
Correction
Last two chapters
How about the opposite problem?
Problem Definition
Proposed Approach - Generator Block Diagram
Text Encoding
Start and End Distributions
Latent Path Construction
Conditional BatchNormalization (CBN)
Frame Generation
UpPooling Block Details
Proposed Approach - Discriminator
Loss Function - Generator
Hinge GAN-Loss on Discriminator
Evaluation Metrics
A2D Quantitative Results
A2D Results
Robotic Results
Dissertation Summary
Future Work
Taught by
UCF CRCV
