Comparative study of camshift and KLT algorithms for real time face detection and tracking applications

Chatterjee, Debmalya; Chandran, Saravanan

doi:10.1109/icrcicn.2016.7813552

Cited by 12 publications

(7 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Object detection and tracking from a video-sequence are becoming an interesting subject in many computer vision applications and artificial intelligence, as bank security, border crossings, airport check-in, home monitoring, office remote meeting, prisons and factories (Barnouti et al, 2018;Chatterjee and Chandran, 2016). The KLT method to allows track a set of feature points in video frames.…”

Section: Kltmentioning

confidence: 99%

“…The KLT method to allows track a set of feature points in video frames. Its computational efficiency and robustness to scale change are relevant aspects that make its use widely feasible in the development of computer vision systems (Barnouti et al, 2018;Chatterjee and Chandran, 2016).…”

Section: Kltmentioning

confidence: 99%

“…The structure of the KLT algorithm is composed of two phases of operation, features extraction and tracking (Chai and Shi, 2011). In the features extraction phase, initially the Viola-Jones algorithm detects the face in the video (Barnouti et al, 2018;Chatterjee and Chandran, 2016) and feature points are identified around of face. With the feature points identified on image, the task of feature tracking is to track them from frame to frame (Chai and Shi, 2011).…”

Section: Kltmentioning

confidence: 99%

“…In different areas, as industrial monitoring, sensor networks, robotics, signal processing, computer vision, pattern recognition, text mining, information security, diagnosis and medical supervision (Pimentel et al, 2014;Oliveira, Moisés A. et al, 2020), studies in the scope of ND has contributed to the development of intelligent systems. In the field of affective computing, natural verbal and non-verbal signals of human emotions, cognitions, perceptions and behaviors are used in the production of efficient and satisfactory interaction systems between man and machine, where the human face is used of several applications (Calvo and D'Mello, 2010;Chatterjee and Chandran, 2016).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Novelty Detection Applied in Recognition of Facial Expressions

Bove

Cerqueira

Filho

2020

Anais Do Congresso Brasileiro De Automática 2020

View full text Add to dashboard Cite

Esta pesquisa investiga a capacidade das redes Perceptron de Múltiplas Camadas (MLP) e Função de Base Radial (RBF) na tarefa de Detecção de Novidade (DN) no reconhecimento de expressões faciais usando recursos de vídeo. O conjunto de dados de vídeo utilizado é produzido por atores profissionais em estúdio com estados afetivos básicos do rosto humano. Os algoritmos Viola-Jones, Kanade-Lucas-Tomasi (KLT) e Análise de Componentes Principais (PCA) são usados na fase de pré-processamento para extrair atributos da face. Os resultados avaliam o desempenho das redes MLP e RBF na tarefa DN, usando novas expressões faciais compatíveis com as utilizadas na fase de treinamento e também examinam a capacidade das redes em DN usando as faces de atores nunca antes vistos pelas redes. Neste processo, as redes MLP e RBF têm uma precisão de 98% para tarefa de classificação, 68% e 98% para DN com dados semelhantes aos dados da fase de treinamento e 100% para DN com dados totalmente novos. Assim, este trabalho reúne métodos e técnicas aplicadas na DN utilizando Redes Neurais Artificiais (RNA), visando a produção de sistemas interativos de cognição no campo da computação afetiva, baseados em técnicas de Inteligência Artificial (IA) e Visão Computacional.

show abstract

Section: Kltmentioning

confidence: 99%

Section: Kltmentioning

confidence: 99%

Section: Kltmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Novelty Detection Applied in Recognition of Facial Expressions

Bove

Cerqueira

Filho

2020

Anais Do Congresso Brasileiro De Automática 2020

View full text Add to dashboard Cite

show abstract

“…The CamSHIFT and KLT algorithms are additional face-tracking mechanisms that have been developed in order to describe faces in a frame-by-frame manner [22]- [25]. CamSHIFT controls color distribution probability by maintaining proper sizes, and uses the meanSHIFT mechanism to locate the center of human faces [23]. On the other hand, the KLT mechanism tracks the movement of a human and is responsible for calculating brightness changes by assuming that every feature in the human face is constant.…”

Section: Related Workmentioning

confidence: 99%

Detection and Tracking of Faces in 3D Using a Stereo Camera Arrangements

AlQahtani¹,

Banks²,

Chandran³

et al. 2019

IJMLC

View full text Add to dashboard Cite

3D facial tracking has become vital to the continued integration of computers, technology, and human society. In recent decades, the integration of technology has increased, and the use of surveillance, conference calls, gaming components, and other similar applications has spurred demand for the ability to recognize the distinctive features of humans. However, in order for these new technologies to function effectively and reach their fullest potential, a great deal of work is still needed. The field of facial mapping and tracking is still in its early developmental stages, necessitating additional research into the best methods of tracking and monitoring specific human faces. To this end, an algorithm has been created that would allow for improvements in this area; however, a video was first required that could be used effectively for the algorithm. Two web cameras running on Raspberry Pi were used to gather the footage necessary for detecting and tracking specific facial features. While certain limitations were identified throughout the process, the algorithm still achieved significant successful tracking results. In spite of this success, further efforts are still needed to effectively explore the proposed algorithm and improve upon these initial results.

show abstract