|
|
Patricia L. Suarez, Angel D. Sappa, & Boris X. Vintimilla. (2017). Colorizing Infrared Images through a Triplet Condictional DCGAN Architecture. In 19th International Conference on Image Analysis and Processing. (pp. 287–297).
|
|
|
|
Patricia L. Suarez, Angel D. Sappa, & Boris X. Vintimilla. (2017). Learning Image Vegetation Index through a Conditional Generative Adversarial Network. In 2nd IEEE Ecuador Tehcnnical Chapters Meeting (ETCM).
|
|
|
|
Milton Mendieta, F. Panchana, B. Andrade, B. Bayot, C. Vaca, Boris X. Vintimilla, et al. (2018). Organ identification on shrimp histological images: A comparative study considering CNN and feature engineering. In IEEE Ecuador Technical Chapters Meeting ETCM 2018. Cuenca, Ecuador (pp. 1–6).
Abstract: The identification of shrimp organs in biology using
histological images is a complex task. Shrimp histological images
poses a big challenge due to their texture and similarity among
classes. Image classification by using feature engineering and
convolutional neural networks (CNN) are suitable methods to
assist biologists when performing organ detection. This work
evaluates the Bag-of-Visual-Words (BOVW) and Pyramid-Bagof-
Words (PBOW) models for image classification leveraging big
data techniques; and transfer learning for the same classification
task by using a pre-trained CNN. A comparative analysis
of these two different techniques is performed, highlighting
the characteristics of both approaches on the shrimp organs
identification problem.
|
|
|
|
Patricia L. Suarez, Angel D. Sappa, Boris X. Vintimilla, & Riad I. Hammoud. (2018). Deep Learning based Single Image Dehazing. In 14th IEEE Workshop on Perception Beyond the Visible Spectrum – In conjunction with CVPR 2018. Salt Lake City, Utah. USA.
Abstract: This paper proposes a novel approach to remove haze
degradations in RGB images using a stacked conditional
Generative Adversarial Network (GAN). It employs a triplet
of GAN to remove the haze on each color channel independently.
A multiple loss functions scheme, applied over a
conditional probabilistic model, is proposed. The proposed
GAN architecture learns to remove the haze, using as conditioned
entrance, the images with haze from which the clear
images will be obtained. Such formulation ensures a fast
model training convergence and a homogeneous model generalization.
Experiments showed that the proposed method
generates high-quality clear images.
|
|
|
|
Patricia L. Suarez, Angel D. Sappa, & Boris X. Vintimilla. (2018). Vegetation Index Estimation from Monospectral Images. In 15th International Conference, Image Analysis and Recognition (ICIAR 2018), Póvoa de Varzim, Portugal. Lecture Notes in Computer Science (Vol. 10882, pp. 353–362).
Abstract: This paper proposes a novel approach to estimate Normalized
Difference Vegetation Index (NDVI) from just the red channel of
a RGB image. The NDVI index is defined as the ratio of the difference
of the red and infrared radiances over their sum. In other words, information
from the red channel of a RGB image and the corresponding
infrared spectral band are required for its computation. In the current
work the NDVI index is estimated just from the red channel by training a
Conditional Generative Adversarial Network (CGAN). The architecture
proposed for the generative network consists of a single level structure,
which combines at the final layer results from convolutional operations
together with the given red channel with Gaussian noise to enhance
details, resulting in a sharp NDVI image. Then, the discriminative model
estimates the probability that the NDVI generated index came from the
training dataset, rather than the index automatically generated. Experimental
results with a large set of real images are provided showing that
a Conditional GAN single level model represents an acceptable approach
to estimate NDVI index.
|
|
|
|
Patricia L. Suarez, Angel D. Sappa, Boris X. Vintimilla, & Riad I. Hammoud. (2018). Near InfraRed Imagery Colorization. In 25 th IEEE International Conference on Image Processing, ICIP 2018 (pp. 2237–2241).
Abstract: This paper proposes a stacked conditional Generative
Adversarial Network-based method for Near InfraRed
(NIR) imagery colorization. We propose a variant architecture
of Generative Adversarial Network (GAN) that uses multiple
loss functions over a conditional probabilistic generative model.
We show that this new architecture/loss-function yields better
generalization and representation of the generated colored IR
images. The proposed approach is evaluated on a large test
dataset and compared to recent state of the art methods using
standard metrics.1
Index Terms—Convolutional Neural Networks (CNN), Generative
Adversarial Network (GAN), Infrared Imagery colorization.
|
|
|
|
Patricia L. Suarez, Angel D. Sappa, & Boris X. Vintimilla. (2018). Adaptive Harris Corners Detector Evaluated with Cross-Spectral Images. In International Conference on Information Technology & Systems (ICITS 2018). ICITS 2018. Advances in Intelligent Systems and Computing (Vol. 721).
Abstract: This paper proposes a novel approach to use cross-spectral
images to achieve a better performance with the proposed Adaptive Harris
corner detector comparing its obtained results with those achieved
with images of the visible spectra. The images of urban, field, old-building
and country category were used for the experiments, given the variety of
the textures present in these images, with which the complexity of the
proposal is much more challenging for its verification. It is a new scope,
which means improving the detection of characteristic points using crossspectral
images (NIR, G, B) and applying pruning techniques, the combination
of channels for this fusion is the one that generates the largest
variance based on the intensity of the merged pixels, therefore, it is that
which maximizes the entropy in the resulting Cross-spectral images.
Harris is one of the most widely used corner detection algorithm, so
any improvement in its efficiency is an important contribution in the
field of computer vision. The experiments conclude that the inclusion of
a (NIR) channel in the image as a result of the combination of the spectra,
greatly improves the corner detection due to better entropy of the
resulting image after the fusion, Therefore the fusion process applied to
the images improves the results obtained in subsequent processes such as
identification of objects or patterns, classification and/or segmentation.
|
|
|
|
Patricia L. Suarez, Angel D. Sappa, & Boris X. Vintimilla. (2018). Cross-spectral image dehaze through a dense stacked conditional GAN based approach. In 14th IEEE International Conference on Signal Image Technology & Internet based Systems (SITIS 2018) (pp. 358–364).
Abstract: This paper proposes a novel approach to remove haze from RGB images using a near infrared images based on a dense stacked conditional Generative Adversarial Network (CGAN). The architecture of the deep network implemented receives, besides the images with haze, its corresponding image in the near infrared spectrum, which serve to accelerate the learning process of the details of the characteristics of the images. The model uses a triplet layer that allows the independence learning of each channel of the visible spectrum image to remove the haze on each color channel separately. A multiple loss function scheme is proposed, which ensures balanced learning between the colors and the structure of the images. Experimental results have shown that the proposed method effectively removes the haze from the images. Additionally, the proposed approach is compared with a state of the art approach showing better results.
|
|
|
|
Miguel Realpe, Boris X. Vintimilla, & L. Vlacic. (2015). Towards Fault Tolerant Perception for autonomous vehicles: Local Fusion. In IEEE 7th International Conference on Cybernetics and Intelligent Systems (CIS) and IEEE Conference on Robotics, Automation and Mechatronics (RAM), Siem Reap, 2015. (pp. 253–258).
Abstract: Many robust sensor fusion strategies have been developed in order to reliably detect the surrounding environments of an autonomous vehicle. However, in real situations there is always the possibility that sensors or other components may fail. Thus, internal modules and sensors need to be monitored to ensure their proper function. This paper introduces a general view of a perception architecture designed to detect and classify obstacles in an autonomous vehicle's environment using a fault tolerant framework, whereas elaborates the object detection and local fusion modules proposed in order to achieve the modularity and real-time process required by the system.
|
|
|
|
Miguel Realpe, Boris X. Vintimilla, & Ljubo Vlacic. (2015). Sensor Fault Detection and Diagnosis for autonomous vehicles. In 2nd International Conference on Mechatronics, Automation and Manufacturing (ICMAM 2015), International Conference on, Singapur, 2015 (Vol. 30, pp. 1–6). EDP Sciences.
Abstract: In recent years testing autonomous vehicles on public roads has become a reality. However, before having autonomous vehicles completely accepted on the roads, they have to demonstrate safe operation and reliable interaction with other traffic participants. Furthermore, in real situations and long term operation, there is always the possibility that diverse components may fail. This paper deals with possible sensor faults by defining a federated sensor data fusion architecture. The proposed architecture is designed to detect obstacles in an autonomous vehicle’s environment while detecting a faulty sensor using SVM models for fault detection and diagnosis. Experimental results using sensor information from the KITTI dataset confirm the feasibility of the proposed architecture to detect soft and hard faults from a particular sensor.
|
|
