scholarly journals Video Face Recognition System for Large Scale Person Re-Identification Using Grassman Algorithm

2019 ◽  
pp. 550-559
Author(s):  
Pandeeshvari. T ◽  
Aajan Kumar
Keyword(s):  
Real Time ◽  
Large Scale ◽  
Recognition System ◽  
Face Image ◽  
Image Resolution ◽  
Video Sequences ◽  
Human Machine Interaction ◽  
Still Images ◽  
Video Frames ◽  
Thermal Face

The identity or verification of humans primarily based on their thermal information isn't always an easy mission to perform, but thermal face biometrics can make contributions to that undertaking. Face reputation is an interesting and a successful application of Image analysis and Pattern recognition. Facial pictures are important for intelligent vision based human machine interaction. Face processing is based at the fact that the records approximately a consumer’s identity may be extracted from the image and the computers can act as a consequence. A thermal face image should be represented with biometrics features that highlight thermal face characteristic and are compact and easy to use for classification. Second, image resolution is basically lower for video sequences. If the subject is present in very far from the camera, the actual face image resolution can be as low as 64 by 64 pixels. Finally, face image variations, such as illumination, expression, pose, occlusion, and motion, are more important in video sequences. The approach can address the unbalanced distributions between still images and videos in a robust way by generating multiple “bridges” to connect the still images and video frames. So in this project, implement still to video matching approach to match the images with videos using Grassmann manifold learning approach to know unknown matches. Finally provide voice alert at the time unknown matching in real time environments. And implement neural network classification algorithms to classify the face images in real time captured videos.

User-Independent Hand Motion Classification With Electromyography

2013 ◽  
Author(s):  
Alison E. Gibson ◽  
Mark R. Ison ◽  
Panagiotis Artemiadis
Keyword(s):  
Real Time ◽  
Large Scale ◽  
High Specificity ◽  
Research Area ◽  
Human Machine Interaction ◽  
Hand Motion ◽  
Intensive Training ◽  
Current State ◽  
Important Research Area ◽  
Leave One Out

Electromyographic (EMG) processing is an important research area with direct applications to prosthetics, exoskeletons and human-machine interaction. Current state of the art decoding methods require intensive training on a single user before it can be utilized, and have been unable to achieve both user-independence and real-time performance. This paper presents a real-time EMG classification method which generalizes across users without requiring an additional training phase. An EMG-embedded sleeve quickly positions and records from EMG surface electrodes on six forearm muscles. An optimized decision tree classifies signals from these sensors into five distinct movements for any given user using EMG energy synergies between muscles. This method was tested on 10 healthy subjects using leave-one-out validation, resulting in an overall accuracy of 79±6.6%, with sensitivity and specificity averaging 66% and 97.6%, respectively, over all classified motions. The high specificity values demonstrate the ability to generalize across users, presenting opportunities for large-scale studies and broader accessibility to EMG-driven applications.

scholarly journals Recognizing Facial Expression using PCA and Genetic Algorithm

2013 ◽  
pp. 117-120
Author(s):  
Mahima Agrawal ◽  
Shubangi. D. Giripunje ◽  
P. R. Bajaj
Keyword(s):  
Genetic Algorithm ◽  
Facial Expression ◽  
Facial Expressions ◽  
Recognition System ◽  
Video Sequences ◽  
Computational Time ◽  
Expression Recognition ◽  
Research Subject ◽  
Still Images ◽  
Recognition Of Facial Expressions

This paper presents an efficient method of recognition of facial expressions in a video. The works proposes highly efficient facial expression recognition system using PCA optimized by Genetic Algorithm .Reduced computational time and comparable efficiency in terms of its ability to recognize correctly are the benchmarks of this work. Video sequences contain more information than still images hence are in the research subject now-a-days and have much more activities during the expression actions. We use PCA, a statistical method to reduce the dimensionality and are used to extract features with the help of covariance analysis to generate Eigen –components of the images. The Eigen-components as a feature input is optimized by Genetic algorithm to reduce the computation cost.

scholarly journals Self-Supervised Video Representation Learning with Space-Time Cubic Puzzles

2019 ◽  
Vol 33 ◽  
pp. 8545-8552 ◽  
Author(s):  
Dahun Kim ◽  
Donghyeon Cho ◽  
In So Kweon
Keyword(s):  
Action Recognition ◽  
Large Scale ◽  
State Of The Art ◽  
Representation Learning ◽  
Space Time ◽  
Temporal Relation ◽  
Still Images ◽  
Video Frames ◽  
Spatio Temporal ◽  
The Cost

Self-supervised tasks such as colorization, inpainting and zigsaw puzzle have been utilized for visual representation learning for still images, when the number of labeled images is limited or absent at all. Recently, this worthwhile stream of study extends to video domain where the cost of human labeling is even more expensive. However, the most of existing methods are still based on 2D CNN architectures that can not directly capture spatio-temporal information for video applications. In this paper, we introduce a new self-supervised task called as Space-Time Cubic Puzzles to train 3D CNNs using large scale video dataset. This task requires a network to arrange permuted 3D spatio-temporal crops. By completing Space-Time Cubic Puzzles, the network learns both spatial appearance and temporal relation of video frames, which is our final goal. In experiments, we demonstrate that our learned 3D representation is well transferred to action recognition tasks, and outperforms state-of-the-art 2D CNN-based competitors on UCF101 and HMDB51 datasets.

scholarly journals Zynq-Based Reconfigurable System for Real-Time Edge Detection of Noisy Video Sequences

2016 ◽  
Vol 2016 ◽  
pp. 1-9 ◽  
Author(s):  
Iljung Yoon ◽  
Heewon Joung ◽  
Jooheung Lee
Keyword(s):  
Edge Detection ◽  
Embedded System ◽  
Real Time ◽  
Video Sequences ◽  
Density Level ◽  
Real Time Processing ◽  
Reconfigurable System ◽  
Video Frames ◽  
Noise Density ◽  
Run Time

We implement Zynq-based self-reconfigurable system to perform real-time edge detection of 1080p video sequences. While object edge detection is a fundamental tool in computer vision, noises in the video frames negatively affect edge detection results significantly. Moreover, due to the high computational complexity of 1080p video filtering operations, hardware implementation on reconfigurable hardware fabric is necessary. Here, the proposed embedded system utilizes dynamic reconfiguration capability of Zynq SoC so that partial reconfiguration of different filter bitstreams is performed during run-time according to the detected noise density level in the incoming video frames. Pratt’s Figure of Merit (PFOM) to evaluate the accuracy of edge detection is analyzed for various noise density levels, and we demonstrate that adaptive run-time reconfiguration of the proposed filter bitstreams significantly increases the accuracy of edge detection results while efficiently providing computing power to support real-time processing of 1080p video frames. Performance results on configuration time, CPU usage, and hardware resource utilization are also compared.

scholarly journals Deep Learning-Based Real-Time Multiple-Person Action Recognition System

Sensors ◽  
2020 ◽  
Vol 20 (17) ◽  
pp. 4758
Author(s):  
Jen-Kai Tsai ◽  
Chen-Chien Hsu ◽  
Wei-Yen Wang ◽  
Shao-Kang Huang
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Action Recognition ◽  
Recognition System ◽  
Video Stream ◽  
Term Care ◽  
Sliding Windows ◽  
Video Frames ◽  
Single Person ◽  
Smart Surveillance

Action recognition has gained great attention in automatic video analysis, greatly reducing the cost of human resources for smart surveillance. Most methods, however, focus on the detection of only one action event for a single person in a well-segmented video, rather than the recognition of multiple actions performed by more than one person at the same time for an untrimmed video. In this paper, we propose a deep learning-based multiple-person action recognition system for use in various real-time smart surveillance applications. By capturing a video stream of the scene, the proposed system can detect and track multiple people appearing in the scene and subsequently recognize their actions. Thanks to high resolution of the video frames, we establish a zoom-in function to obtain more satisfactory action recognition results when people in the scene become too far from the camera. To further improve the accuracy, recognition results from inflated 3D ConvNet (I3D) with multiple sliding windows are processed by a nonmaximum suppression (NMS) approach to obtain a more robust decision. Experimental results show that the proposed method can perform multiple-person action recognition in real time suitable for applications such as long-term care environments.

A BACK PROPAGATION BASED REAL-TIME LICENSE PLATE RECOGNITION SYSTEM

2008 ◽  
Vol 22 (02) ◽  
pp. 233-251 ◽  
Author(s):  
YO-PING HUANG ◽  
TSUN-WEI CHANG ◽  
YEN-REN CHEN ◽  
FRODE EIKA SANDNES
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Recognition Rate ◽  
Back Propagation ◽  
Recognition System ◽  
Back Propagation Neural Network ◽  
License Plate ◽  
License Plate Recognition ◽  
Video Frames ◽  
Parking Lot

License plate recognition systems have been used extensively for many applications including parking lot management, tollgate monitoring, and for the investigation of stolen vehicles. Most researches focus on static systems, which require a clear and level image to be taken of the license plate. However, the acquisition of images that can be successfully analyzed relies on both the location and movement of the target vehicle and the clarity of the environment. Moreover, only few studies have addressed the problems associated with instant car image processing. In view of these problems, a real-time license plate recognition system is proposed that recognizes the video frames taken from existing surveillance cameras. The proposed system finds the location of the license plate using projection analysis, and the characters are identified using a back propagation neural network. The strategy achieves a recognition rate of 85.8% and almost 100% after the neural network has been retrained using the erroneously recognized characters, respectively.

scholarly journals MS-Faster R-CNN: Multi-Stream Backbone for Improved Faster R-CNN Object Detection and Aerial Tracking from UAV Images

2021 ◽  
Vol 13 (9) ◽  
pp. 1670
Author(s):  
Danilo Avola ◽  
Luigi Cinque ◽  
Anxhelo Diko ◽  
Alessio Fagioli ◽  
Gian Luca Foresti ◽  
...  
Keyword(s):  
Image Analysis ◽  
Object Detection ◽  
Real Time ◽  
Video Sequences ◽  
Multi Scale ◽  
Video Frames ◽  
Common Strategy ◽  
Real Time Tracking ◽  
Uav Images ◽  
Sort Algorithm

Tracking objects across multiple video frames is a challenging task due to several difficult issues such as occlusions, background clutter, lighting as well as object and camera view-point variations, which directly affect the object detection. These aspects are even more emphasized when analyzing unmanned aerial vehicles (UAV) based images, where the vehicle movement can also impact the image quality. A common strategy employed to address these issues is to analyze the input images at different scales to obtain as much information as possible to correctly detect and track the objects across video sequences. Following this rationale, in this paper, we introduce a simple yet effective novel multi-stream (MS) architecture, where different kernel sizes are applied to each stream to simulate a multi-scale image analysis. The proposed architecture is then used as backbone for the well-known Faster-R-CNN pipeline, defining a MS-Faster R-CNN object detector that consistently detects objects in video sequences. Subsequently, this detector is jointly used with the Simple Online and Real-time Tracking with a Deep Association Metric (Deep SORT) algorithm to achieve real-time tracking capabilities on UAV images. To assess the presented architecture, extensive experiments were performed on the UMCD, UAVDT, UAV20L, and UAV123 datasets. The presented pipeline achieved state-of-the-art performance, confirming that the proposed multi-stream method can correctly emulate the robust multi-scale image analysis paradigm.

scholarly journals Real-Time Mask Recognition

IoT ◽  
2021 ◽  
Vol 2 (4) ◽  
pp. 688-716
Author(s):  
Rachel M. Billings ◽  
Alan J. Michaels
Keyword(s):  
Neural Network ◽  
Low Power ◽  
Real Time ◽  
Low Cost ◽  
Model Performance ◽  
Network Models ◽  
Recognition System ◽  
Neural Network Models ◽  
Still Images ◽  
Experimental Dataset

While a variety of image processing studies have been performed to quantify the potential performance of neural network-based models using high-quality still images, relatively few studies seek to apply those models to a real-time operational context. This paper seeks to extend prior work in neural-network-based mask detection algorithms to a real-time, low-power deployable context that is conducive to immediate installation and use. Particularly relevant in the COVID-19 era with varying rules on mask mandates, this work applies two neural network models to inference of mask detection in both live (mobile) and recorded scenarios. Furthermore, an experimental dataset was collected where individuals were encouraged to use presentation attacks against the algorithm to quantify how perturbations negatively impact model performance. The results from evaluation on the experimental dataset are further investigated to identify the degradation caused by poor lighting and image quality, as well as to test for biases within certain demographics such as gender and ethnicity. In aggregate, this work validates the immediate feasibility of a low-power and low-cost real-time mask recognition system.

Sign in / Sign up

Export Citation Format

Share Document