Video+Language: From Classification to Description

Video + Language
Jiebo Luo
Department of Computer Science

Introduction
• Video has become ubiquitous on the Internet, TV, as well as
personal devices.
• Recognition of video content has been a fundamental challenge
in computer vision for decades, where previous research
predominantly focused on understanding videos using a
predefined yet limited vocabulary.
• Thanks to the recent development of deep learning techniques,
researchers in both computer vision and multimedia
communities are now striving to bridge video with
natural language, which can be regarded as the ultimate goal of
video understanding.
• We present recent advances in exploring the synergy of video
understanding and language processing, including video-
language alignment, video captioning, and video emotion
analysis.

From Classification to Description
Recognizing Realistic Actions from Videos "in the Wild"
UCF-11 to UCF-101
(CVPR 2009)
Similarity btw Videos Cross-Domain Learning
Visual Event Recognition in Videos by
Learning from Web Data
(CVPR2010 Best Student Paper)
Heterogeneous Feature Machine For Visual Recognition
(ICCV 2009)

4
Semantic Video Entity Linking
(ICCV2015)

Exploring Coherent Motion Patterns
via Structured Trajectory Learning for
Crowd Mood Modeling
(IEEE T-CSVT 2016)

Aligning Language Descriptions
with Videos
Iftekhar Naim, Young Chol Song, Qiguang Liu
Jiebo Luo, Dan Gildea, Henry Kautz
link

Unsupervised Alignment of Actions
in Video with Text Descriptions
Y. Song, I. Naim, A. Mamun, K. Kulkarni, P. Singla
J. Luo, D. Gildea, H. Kautz

OverviewOverview
 Unsupervised alignment of video with text
 Motivations
 Generate labels from data (reduce burden of manual labeling)
 Learn new actions from only parallel video+text
 Extend noun/object matching to verbs and actions
 Unsupervised alignment of video with text
 Motivations
 Generate labels from data (reduce burden of manual labeling)
 Learn new actions from only parallel video+text
 Extend noun/object matching to verbs and actions
Matching Verbs to Actions
The person takes out a knife
and cutting board
Matching Nouns to Objects
[Naim et al., 2015]
An overview of the text and
video alignment framework

Hyperfeatures for ActionsHyperfeatures for Actions
 High-level features required for alignment with text
→ Motion features are generally low-level
 Hyperfeatures, originally used for image recognition extended
for use with motion features
→ Use temporal domain instead of spatial domain for vector
quantization (clustering)
 High-level features required for alignment with text
→ Motion features are generally low-level
 Hyperfeatures, originally used for image recognition extended
for use with motion features
→ Use temporal domain instead of spatial domain for vector
quantization (clustering)
Originally described in “Hyperfeatures:
Multilevel Local Coding for Visual Recog‐
nition” Agarwal, A. (ECCV 06), for images Hyperfeatures for actions

Hyperfeatures for ActionsHyperfeatures for Actions
 From low-level motion features, create high-level
representations that can easily align with verbs in text
 From low-level motion features, create high-level
representations that can easily align with verbs in text
Cluster 3
at time t
Accumulate over
frame at time t
& cluster
Conduct vector
quantization
of the histogram
at time t
Cluster 3, 5, …,5,20
= Hyperfeature 6
Each color code
is a vector
quantized
STIP point
Vector quantized
STIP point histogram at time t
Accumulate clusters over
window (t‐w/2, t+w/2]
and conduct vector
quantization
→ first‐level hyperfeatures
Align hyperfeatures
with verbs from text
(using LCRF)

Latent-variable CRF AlignmentLatent-variable CRF Alignment
 CRF where the latent variable is the alignment
 N pairs of video/text observations {(xi, yi)}i=1 (indexed by i)
 Xi,m represents nouns and verbs extracted from the mth sentence
 Yi,n represents blobs and actions in interval n in the video
 Conditional likelihood
 conditional probability of
 Learning weights w
 Stochastic gradient descent
 CRF where the latent variable is the alignment
 N pairs of video/text observations {(xi, yi)}i=1 (indexed by i)
 Xi,m represents nouns and verbs extracted from the mth sentence
 Yi,n represents blobs and actions in interval n in the video
 Conditional likelihood
 conditional probability of
 Learning weights w
 Stochastic gradient descent
where feature function
More details in Naim et al. 2015 NAACL Paper ‐
Discriminative unsupervised alignment of natural language instructions with corresponding video segments
N

Experiments: Wetlab DatasetExperiments: Wetlab Dataset
 RGB-Depth video with lab protocols in text
 Compare addition of hyperfeatures generated from motion
features to previous results (Naim et al. 2015)
 Small improvement over previous results
 Activities already highly correlated with object-use
 RGB-Depth video with lab protocols in text
 Compare addition of hyperfeatures generated from motion
features to previous results (Naim et al. 2015)
 Small improvement over previous results
 Activities already highly correlated with object-use
Detection of objects in 3D space
using color and point‐cloud
Previous results
using object/noun
alignment only
Addition of different types
of motion features
2DTraj: Dense trajectories
*Using hyperfeature window size w=150

Experiments: TACoS DatasetExperiments: TACoS Dataset
 RGB video with crowd-sourced text descriptions
 Activities such as “making a salad,” “baking a cake”
 No object recognition, alignment using actions only
 Uniform: Assume each sentence takes the same amount of time over the entire sequence
 Segmented LCRF: Assume the segmentation of actions is known, infer only the action labels
 Unsupervised LCRF: Both segmentation and alignment are unknown
 Effect of window size and number of clusters
 Consistent with average
action length: 150 frames
 RGB video with crowd-sourced text descriptions
 Activities such as “making a salad,” “baking a cake”
 No object recognition, alignment using actions only
 Uniform: Assume each sentence takes the same amount of time over the entire sequence
 Segmented LCRF: Assume the segmentation of actions is known, infer only the action labels
 Unsupervised LCRF: Both segmentation and alignment are unknown
 Effect of window size and number of clusters
 Consistent with average
action length: 150 frames
*Using hyperfeature
window size w=150
*d(2)=64

Experiments: TACoS DatasetExperiments: TACoS Dataset
 Segmentation from a sequence in the dataset Segmentation from a sequence in the dataset
Crowd‐sourced descriptions
Example of text and video alignment generated
by the system on the TACoS corpus for sequence s13‐d28

Image Captioning with Semantic
Attention (CVPR 2016)
Quanzeng You, Jiebo Luo
Hailin Jin, Zhaowen Wang and Chen Fang

Image Captioning
• Motivations
– Real-world Usability
• Help visually impaired people, learning-impaired
– Improving Image Understanding
• Classification, Objection detection
– Image Retrieval
1. a young girl inhales with the intent of blowing
out a candle
2. girl blowing out the candle on an ice cream
1. A shot from behind home
plate of children playing
baseball
2. A group of children playing
baseball in the rain
3. Group of baseball players
playing on a wet field

Introduction of Image Captioning
• Machine learning as an approach to solve the problem

Overview
• Brief overview of current approaches
• Our main motivation
• The proposed semantic attention model
• Evaluation results

Brief Introduction of Recurrent Neural Network
• Different from CNN
11),(   ttttt BhAxhxfh
tt Chy 
• Unfolding over time
 Feedforward network
 Backpropagation through time
Inputs
Hidden Units
Outputs
xt ht-1
ht
yt
Inputs
Hidden Units
Outputs
C
yt
Inputs
Hidden Units
Inputs
Hidden Units
B
B
A
A
A B
t-1
t-2

Applications of Recurrent Neural Networks
• Machine Translation
• Reads input sentence “ABC” and produces “WXYZ”
Decoder RNNEncoder RNN

Encoder-Decoder Framework for Captioning
• Inspired by neural network based machine
translation
• Loss function



N
t
tt wwIwp
IwpL
1
10 ),,,|(log
)|(log


Our Motivation
• Additional textual information
– Own noisy title, tags or captions (Web)

Our Motivation
– Visually similar nearest neighbor images

Our Motivation
– Visually similar nearest neighbor images
– Success of low-level tasks
• Visual attributes detection

Image Captioning with Semantic Attention
• Big idea

First Idea
• Provide additional knowledge at each input node
• Concatenate the input word and the extra attributes K
• Each image has a fixed keyword list
)],,([),( 11   tktttt hbKWwfhxfh
Visual Features: 1024
GoogleNet
LSTM Hidden states: 512
Training details:
1. 256 image/sentence
pairs
2. RMS-Prob

Using Attributes along with Visual Features
• Provide additional knowledge at each input node
• Concatenate the visual embedding and keywords for h0
];[),( 10 bKWvWhvfh kiv  

Attention Model on Attributes
• Instead of using the same set of attributes at every
step
• At each step, select the attributes (attention)
 m mtmt kKwatt ),(
)softmax VK(wT
tt 
))],,(;([),( 11   tttttt hKwattxfhxfh

Overall Framework
• Training with a bilinear/bilateral attention model
ht
pt
xt
v
{Ai}
Yt~
RNN
Image
CNN
AttrDet 1
AttrDet 2
AttrDet 3
AttrDet N

t = 0
Word

Visual Attributes
• A secondary contribution
• We try different approaches

Performance
• Examples showing the impact of visual attributes on captions

Performance on the Testing Dataset
• Publicly available split

Performance
• MS-COCO Image Captioning Challenge

TGIF: A New Dataset and Benchmark on
Animated GIF Description
Yuncheng Li, Yale Song, Liangliang Cao, Joel Tetreault,
Larry Goldberg, Jiebo Luo

Comparison with Existing Datasets

Examples
a skate boarder is
doing trick on his skate
board.
a gloved hand opens to
reveal a golden ring.
a sport car is swinging on
the race playground
the vehicle is moving fast
into the tunnel

Contributions
• A large scale animated GIF description dataset for promoting image
sequence modeling and research
• Performing automatic validation to collect natural language descriptions
from crowd workers
• Establishing baseline image captioning methods for future
benchmarking
• Comparison with existing datasets, highlighting the benefits with
animated GIFs

In Comparison with Existing Datasets
• The language in our dataset is closer to common language
• Our dataset has an emphasis on the verbs
• Animated GIFs are more coherent and self contained
• Our dataset can be used to solve more difficult movie description
problem

Machine Generated Sentence Examples

Comparing Professionals and Crowd-workers
Crowd worker: two people are kissing on a boat.
Professional: someone glances at a kissing couple
then steps to a railing overlooking the ocean an
older man and woman stand beside him.
Crowd worker: two men got into their car
and not able to go anywhere because the
wheels were locked.
Professional: someone slides over the
camaros hood then gets in with his partner
he starts the engine the revving vintage
car starts to backup then lurches to a halt.
Crowd worker: a man in a shirt and tie sits beside a person who is
covered in a sheet.
Professional: he makes eye contact with the woman for only a second.
More: http://beta-
web2.cloudapp.net/ls
mdc_sentence_comp
arison.html

Movie Descriptions versus TGIF
• Crowd workers are encouraged to describe the major visual content
directly, and not to use overly descriptive language
• Because our animated GIFs are presented to crowd workers without
any context, the sentences in our dataset are more self-contained
• Animated GIFs are perfectly segmented since they are carefully
curated by online users to create a coherent visual story

Code & Dataset
• Yahoo! webscope (Coming soon!)
• Animated GIFs and sentences
• Code and models for LSTM baseline
• Pipeline for syntactic and semantic validation to collect natural
languages from crowd workers

Thanks
Q & A
Visual Intelligence & Social Multimedia Analytics
Google
***
Baidu
Sogou
Bing
XiaoIce
How Intelligent Are the AI Systems Today?

Video+Language: From Classification to Description

More Related Content

What's hot (19)

Similar to Video+Language: From Classification to Description (20)

More from Goergen Institute for Data Science (9)

Recently uploaded (20)

Video+Language: From Classification to Description