Identifying facemask-wearing condition using image super-resolution with classification network to prevent COVID-19
The rapid worldwide spread of Coronavirus Disease 2019 (COVID-19) has resulted in a
global pandemic. Correct facemask wearing is valuable for infectious disease control, but …
global pandemic. Correct facemask wearing is valuable for infectious disease control, but …
Appearance descriptors for person re-identification: a comprehensive review
R Satta - arxiv preprint arxiv:1307.5748, 2013 - arxiv.org
In video-surveillance, person re-identification is the task of recognising whether an
individual has already been observed over a network of cameras. Typically, this is achieved …
individual has already been observed over a network of cameras. Typically, this is achieved …
Real-time illumination invariant lane detection for lane departure warning system
J Son, H Yoo, S Kim, K Sohn - Expert Systems with Applications, 2015 - Elsevier
Lane detection is an important element in improving driving safety. In this paper, we propose
a real-time and illumination invariant lane detection method for lane departure warning …
a real-time and illumination invariant lane detection method for lane departure warning …
Color invariants for person reidentification
We revisit the problem of specific object recognition using color distributions. In some
applications-such as specific person identification-it is highly likely that the color distributions …
applications-such as specific person identification-it is highly likely that the color distributions …
Person re-identification using spatial covariance regions of human body parts
In many surveillance systems there is a requirement to determine whether a given person of
interest has already been observed over a network of cameras. This is the person re …
interest has already been observed over a network of cameras. This is the person re …
One-shot metric learning for person re-identification
Re-identification of people in surveillance footage must cope with drastic variations in color,
background, viewing angle and a person's pose. Supervised techniques are often the most …
background, viewing angle and a person's pose. Supervised techniques are often the most …
A two-stage data-driven approach for image-based wind turbine blade crack inspections
A two-stage approach for precisely detecting wind turbine blade surface cracks via
analyzing blade images captured by unmanned aerial vehicles (UAVs) is proposed in this …
analyzing blade images captured by unmanned aerial vehicles (UAVs) is proposed in this …
Automated inspection of PCB components using a genetic algorithm template-matching approach
AJ Crispin, V Rankov - The International Journal of Advanced …, 2007 - Springer
Automated inspection of surface mount PCB boards is a requirement to assure quality and to
reduce manufacturing scrap costs and rework. This paper investigates methodologies for …
reduce manufacturing scrap costs and rework. This paper investigates methodologies for …
Influence of normalization and color space to color texture classification
Color texture classification has recently attracted significant attention due to its multiple
applications. The color texture images depend on the texture surface and its albedo, the …
applications. The color texture images depend on the texture surface and its albedo, the …
Multiple-shot human re-identification by mean riemannian covariance grid
Human re-identification is defined as a requirement to determine whether a given individual
has already appeared over a network of cameras. This problem is particularly hard by …
has already appeared over a network of cameras. This problem is particularly hard by …