Abstract:As image processing techniques and devices advance, the real-time applications of computer vision such as human action/interaction recognition and video content analysis become more attractive. However, the methods proposed in the state-of-the-art studies are still far from representing real-time and all-inclusive classifiers because of the image and video analysis complexity. This work presents a new approach based on key-poses of frame silhouettes for human interaction recognition. We use an inner-distance-based shape descriptor which gives a perfect description of the shape due to its ability to collect data from the whole shape. The core idea is to develop a two-step classifier based on a sequential pattern mining classifier. So, we extract the Bilateral Silhouette shape for the persons and describe it based on the inner-distance feature to compare each frame with a pre-defined dictionary of key-poses. The classification process is performed in frame and sequence layers. Accurate and efficient, the sequential pattern mining approach provides an appealing solution to the problem of sequence classification, giving comparable or even better results than standard classifiers. We evaluated the recognition performance of the system using video sequences of SBU human interaction dataset and the UT-interaction dataset as two well-known interaction datasets and the results are considered acceptable (95.25% in SBU and 90.5% in UT databases, respectively), outperforming most state-of-the-art results. These recognition rates are calculated after we have tested different parameters which can affect the results. Both datasets include multiple interaction classes performed by different actors, which helps us develop an all-inclusive method based on the datasets. The proposed method can be optimized to be used in some real world applications such as abnormal activity recognition in crowded places, auxiliary surveillance system, human-computer interaction, etc.

SneakPeek: Interest Mining of Images based on User Interaction

ShotVis: Smartphone-Based Visualization of OCR Information from Images

Learning User Interest for Image Browsing on Small-Form-factor Devices

FocalClick: Towards Practical Interactive Image Segmentation.

Capturing Social Cues with Imaging Glasses.

Multimodal analysis of user behavior and browsed content under different image search intents

A Novel Hybrid BCI Web Browser Based on SSVEP and Eye-Tracking.

Adaptive user interface of product recommendation based on eye-tracking

A CNN based real-time eye tracker for web mining applications

DEEP: 3D Gaze Pointing in Virtual Reality Leveraging Eyelid Movement

How Deep Is Your Gaze? Leveraging Distance in Image-Based Gaze Analysis

SRPeek: Super Resolution Enabled Screen Peeking Via COTS Smartphone.

Continuous Prediction of Web User Visual Attention on short span Windows based on Gaze Data Analytics

Efficient Online Engagement Analytics Algorithm Toolkit That Can Run on Edge

DynamicRead: Exploring Robust Gaze Interaction Methods for Reading on Handheld Mobile Devices under Dynamic Conditions

Applying Visual User Interest Profiles for Recommendation and Personalisation.

Accelerating eye movement research via accurate and affordable smartphone eye tracking

Pull and Push: Proximity-Aware User Interface for Navigating in 3D Space Using a Handheld Camera

Predicting Eye Gaze Location on Websites

DialPlate: Enhancing the Detection of Smooth Pursuits Eye Movements Using Linear Regression

Two-person interaction recognition using a two-step sequential pattern classification