mining frequent events from video

14
MINING FREQUENT EVENTS FROM VIDEO Steffi Keran Rani J M.E. Multimedia Technology Anna University

Upload: steffi-keran-rani-j

Post on 10-Apr-2017

47 views

Category:

Engineering


1 download

TRANSCRIPT

Page 1: Mining Frequent Events From Video

MINING FREQUENT EVENTS FROM VIDEO

Steffi Keran Rani JM.E. Multimedia Technology

Anna University

Page 2: Mining Frequent Events From Video

EVENT DETECTION Event detection involves the automatic organization of a

multimedia collection C into groups of items, each (group) of which corresponds to a distinct event.

Page 3: Mining Frequent Events From Video

CHALLENGES1. requires application of several

Computer Vision

2. Involves subtleties that are readily

understood by humans, difficult to

encode for machine learning

approaches

3. Can be complicated due to clutter

in the environment, lighting, camera

placement, traffic, etc.

Page 4: Mining Frequent Events From Video

APPLICATIONS

1. Video Surveillance

2. Video- on- Demand

3. Broadcast Video

4. Web Search

Page 5: Mining Frequent Events From Video

CLUSTER CLASSIFICATION#

user

s /

#ph

otos

duration

[1 day, 2 users / 10 photos]

[2 years, 50 users / 120 photos]

#5

LANDMARK

EVENT

Page 6: Mining Frequent Events From Video

EVENT DETECTION USING DATA MINING TECHNIQUES

Video

Video Parsing and Feature Detection

Instance Self Learning

Filtering and Reconstruction

Self Refining Training Dataset

Final DetectionDecision Tree Model

Page 7: Mining Frequent Events From Video

VIDEO PARSING

Page 8: Mining Frequent Events From Video

3 BUILDING BLOCKS1. Video Parsing and Feature Extraction

Involves temporal partitioning of the video sequence into meaningful units.

This module computes a large array of multimodal features (both visual and audio) from input videos

Five visual features are extracted for each shot:

1. pixel_change 2. histo_change;

3. background_mean 4. background_varr 5. dominant_color_ratio

2. Base ClassifiersMultiple base classifiers independently compute detection scores based on available features

3. Score FusionThis module combines multiple base classifier scores through diverse fusion methods, and

computes a single final detection score per video clip

Page 9: Mining Frequent Events From Video

TWO- STEP PROCEDURE

1. Video content processing: The video clip is segmented into certain analysis units and their representative features are extracted.

2. Decision making: process that extracts the semantic index from the feature descriptors.

Page 10: Mining Frequent Events From Video

DECISION MAKING PROCESSDECISION MAKING

Knowledge Based Approaches

Rule based Classifier

Statistical Approaches

Support Vector Machines

Dynamic Bayesian Network

C4.5 decision trees

Page 11: Mining Frequent Events From Video

11

1. Event Detection Using Multi Modal Feature Fusion

Page 12: Mining Frequent Events From Video

2. VIDEO EVENT DETECTION BY INFERRING TEMPORAL INSTANCE LABELS

Video recognition algorithm is inspired by proportion SVM (p-SVM), which explicitly models the latent unknown instance labels together with the known label proportions in a large-margin framework

Page 13: Mining Frequent Events From Video
Page 14: Mining Frequent Events From Video