Robust Person Tracking Algorithm Based on Convolutional Neural Network for Indoor Video Surveillance Systems

Abstract

In this paper, we present an algorithm for multi person tracking in indoor surveillance systems based on tracking-by-detection approach. Convolutional Neural Networks (CNNs) for detection and tracking both are used. CNN Yolov3 has been utilized as detector. Person features extraction is performed based on modified CNN ResNet. Proposed architecture includes 29 convolutional and one fully connected layer. Hungarian algorithm is applied for objects association. After that object visibility in the frame is determined based on CNN and color features. For algorithm evaluation prepared videos that was labeled and tested using MOT evaluation metric. The proposed algorithm efficiency is illustrated and confirmed by our experimental results.

Enhancing spatial perception through sound: mapping human movements into MIDI

“…They use two challenging video surveillance datasets, such as PETS2009 and UMN crowd analysis datasets, to demonstrate their proposed system's effectiveness, which achieved 88.7% and 95.5% of accuracy and detection rate, respectively. In [4] authors propose an algorithm for multi-person tracking in indoor surveillance systems based on a tracking-by-detection approach. They use Convolutional Neural Networks (CNNs) for detecting and tracking people.…”

“…In [23], authors define a lightweight tracking algorithm named Kerman (Kernelized Kalman filter), which is a decision tree based hybrid Kernelized Correlation Filter (KCF) algorithm for human object tracking. In [4,23,30] several machine learning models are used to detect people during video surveillance activities, showing results in terms of classification achieved.…”

“…In the state of the art, it is possible to find different tracking techniques for detecting human movements [4], and techniques to represent their semantics [1]. They have been mainly applied to scenarios in which it is necessary to support humans in the real-time interpretation of critical situations of daily life, such as in the context of video surveillance [7].…”

Perceiving space through sound: mapping human movements into MIDI

“…They use two challenging video surveillance datasets, such as PETS2009 and UMN crowd analysis datasets, to demonstrate the effectiveness of their proposed system, which achieved 88.7% and 95.5% of accuracy and detection rate, respectively. In [4] authors propose an algorithm for multi-person tracking in indoor surveillance systems based on a tracking-bydetection approach. They use Convolutional Neural Networks (CNNs) for detecting and tracking people.…”

Robust object detection in images corrupted by impulse noise

“…Even in solving only one task uncertainty may also occur. For example, to correctly identity tracked people in a current frame, a maximum accuracy of the estimation for people feature similarity with previous frames is required [12]. However, high accuracy for high density noise or overlapping parts of people by objects will lead to loss their index.…”