Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

Human Activity Recognition

•
•
A
AshwinGill1

Human Activity Recognition (HAR) systems aim to recognize human activities through sensors in order to provide assistance. The key steps in designing a HAR system are: 1) Acquiring sensor data and preprocessing it by removing noise. 2) Segmenting the preprocessed data into windows that may contain activities. 3) Extracting features from each window to reduce the data into discriminative features. 4) Training a classification model on the extracted features to predict activity labels, and evaluating the model's performance using methods like a confusion matrix.

1 of 36
Download to read offline
Human Activity Recognition (HAR) - Ashwin Singh (2017222)
Introduction The goal of HAR is to provide information on a user’s behaviour that allows computing systems to proactively assist users with their tasks. Why do we need HAR?
Current Research Computer Vision (Stationary Settings) Inertial Sensors (Dynamic Settings) On-body sensing goes beyond instrumented rooms to provide smart assistance anywhere and at any time.
MAGIC : A Motion Gesture Design Tool
LiveNet - MiT Media Labs
Human Activity Recognition
Themes ● ● ● ● ●
Human Activity Recognition
NIKE+
How would you go about designing a HAR System for you project? ● ARC (Activity Recognition Chain) ○ Sensor Data Acquisition ○ Preprocessing ○ Data Segmentation ○ Feature Extraction and Selection ○ Training + Classification ○ Evaluation ● Challenges you might face
● ● ● ●
Activity Recognition Chain (ARC)
Hmm… such math much wow
Angel wants to build an HAR system. ● ● ● ●
Sensor Data Acquisition 1/6 Sensor Output Format Si = (d1 , d2 ,... dj , …., dt ) i : ith sensor | j : output of sensor at time t = j ● dj ’s can be multidimensional. ● Each sensor is sampled at regular intervals. However, sampling rates of different types of sensors might differ.
Factors she takes into account while sensing
What does the sensor data look like?
Preprocessing 2/6 Sensor data Preprocessing Preprocessed Time Series D/ (synchronization, noise removal) n: total data dimensions, t: # of samples. ● Generic, only dependent on data. ● Algorithms which preserve relevant signal characteristics, only filter noise.
How can she make her data usable to extract features?
Data Segmentation 3/6 To identify time-based segments of preprocessed data that may contain relevant information. D/ W = { w1 , w2 , …, wi , …, wm } | wi = (tstart , tend ) Different Techniques ● Sliding Window ● Energy Based ● Additional Sensors, Contextual Sources
What features can she extract from the Data? ● ● ● ● ●
Feature Extraction & Selection 4/6 Reducing signals into features that are discriminative of the activity. Feature Vector Xi = F(D/ ,wi ) | Xi : [ ]n x 1 where n: # of features Features ● Corresponding to the same activity should be clustered together. ● Can be based on ○ Signals, Body-Model ● There are built-in methods for feature ranking, selection.
More features More dimensions in data Trade-offs: memory, computational power
How can she build a model using the data which classifies live data streams into activities?
Training 5/6 Training Data T = { (Xi , yi ) }i = 1 to N | Xi : Feature Vector, yi : Label Goal : To minimize classification error on T. Training Method ● an algorithm which trains the model to predict label yi , given a feature vector Xi ● Selected on the basis of ○ Type of Activity ○ Complexity of Feature Space
Classification 5.5/6 A two-step process 1. Each Xi is assigned confidence values w.r.t class labels yi ’s. 2. The class label with the highest confidence value is chosen. If all confidence values are below a certain threshold, a null class label is assigned to the corresponding data sample.
Different Types of Classifiers
After selecting the best model for her use-case, how can she evaluate its performance?
Performance Evaluation 6/6 How robust is your designed HAR system? ● Confusion Matrix ● (True/False)-(Positive/Negatives) ● Precision ● Recall
Confusion Matrix
Human Activity Recognition
Human Activity Recognition
Human Activity Recognition
Common Challenges 1/2 ● Intraclass Variability ● Interclass Similarity ● Null Class Problem
HAR Specific Challenges 2/2 ● Definition + Diversity of Human Activities ● Class Imbalance ● Ground Truth Annotation
Human Activity Recognition

More Related Content

Human Activity Recognition

  • 1. Human Activity Recognition (HAR) - Ashwin Singh (2017222)
  • 2. Introduction The goal of HAR is to provide information on a user’s behaviour that allows computing systems to proactively assist users with their tasks. Why do we need HAR?
  • 3. Current Research Computer Vision (Stationary Settings) Inertial Sensors (Dynamic Settings) On-body sensing goes beyond instrumented rooms to provide smart assistance anywhere and at any time.
  • 4. MAGIC : A Motion Gesture Design Tool
  • 5. LiveNet - MiT Media Labs
  • 10. How would you go about designing a HAR System for you project? ● ARC (Activity Recognition Chain) ○ Sensor Data Acquisition ○ Preprocessing ○ Data Segmentation ○ Feature Extraction and Selection ○ Training + Classification ○ Evaluation ● Challenges you might face
  • 13. Hmm… such math much wow
  • 14. Angel wants to build an HAR system. ● ● ● ●
  • 15. Sensor Data Acquisition 1/6 Sensor Output Format Si = (d1 , d2 ,... dj , …., dt ) i : ith sensor | j : output of sensor at time t = j ● dj ’s can be multidimensional. ● Each sensor is sampled at regular intervals. However, sampling rates of different types of sensors might differ.
  • 16. Factors she takes into account while sensing
  • 17. What does the sensor data look like?
  • 18. Preprocessing 2/6 Sensor data Preprocessing Preprocessed Time Series D/ (synchronization, noise removal) n: total data dimensions, t: # of samples. ● Generic, only dependent on data. ● Algorithms which preserve relevant signal characteristics, only filter noise.
  • 19. How can she make her data usable to extract features?
  • 20. Data Segmentation 3/6 To identify time-based segments of preprocessed data that may contain relevant information. D/ W = { w1 , w2 , …, wi , …, wm } | wi = (tstart , tend ) Different Techniques ● Sliding Window ● Energy Based ● Additional Sensors, Contextual Sources
  • 21. What features can she extract from the Data? ● ● ● ● ●
  • 22. Feature Extraction & Selection 4/6 Reducing signals into features that are discriminative of the activity. Feature Vector Xi = F(D/ ,wi ) | Xi : [ ]n x 1 where n: # of features Features ● Corresponding to the same activity should be clustered together. ● Can be based on ○ Signals, Body-Model ● There are built-in methods for feature ranking, selection.
  • 23. More features More dimensions in data Trade-offs: memory, computational power
  • 24. How can she build a model using the data which classifies live data streams into activities?
  • 25. Training 5/6 Training Data T = { (Xi , yi ) }i = 1 to N | Xi : Feature Vector, yi : Label Goal : To minimize classification error on T. Training Method ● an algorithm which trains the model to predict label yi , given a feature vector Xi ● Selected on the basis of ○ Type of Activity ○ Complexity of Feature Space
  • 26. Classification 5.5/6 A two-step process 1. Each Xi is assigned confidence values w.r.t class labels yi ’s. 2. The class label with the highest confidence value is chosen. If all confidence values are below a certain threshold, a null class label is assigned to the corresponding data sample.
  • 27. Different Types of Classifiers
  • 28. After selecting the best model for her use-case, how can she evaluate its performance?
  • 29. Performance Evaluation 6/6 How robust is your designed HAR system? ● Confusion Matrix ● (True/False)-(Positive/Negatives) ● Precision ● Recall
  • 34. Common Challenges 1/2 ● Intraclass Variability ● Interclass Similarity ● Null Class Problem
  • 35. HAR Specific Challenges 2/2 ● Definition + Diversity of Human Activities ● Class Imbalance ● Ground Truth Annotation