Cristhian A. Aguilera, C. A., Cristóbal A. Navarro, & Angel D. Sappa. (2020). Fast CNN Stereo Depth Estimation through Embedded GPU Devices. Sensors 2020, Vol. 2020-June(11), pp. 1–13.
Abstract: Current CNN-based stereo depth estimation models can barely run under real-time
constraints on embedded graphic processing unit (GPU) devices. Moreover, state-of-the-art
evaluations usually do not consider model optimization techniques, being that it is unknown what is
the current potential on embedded GPU devices. In this work, we evaluate two state-of-the-art models
on three different embedded GPU devices, with and without optimization methods, presenting
performance results that illustrate the actual capabilities of embedded GPU devices for stereo depth
estimation. More importantly, based on our evaluation, we propose the use of a U-Net like architecture
for postprocessing the cost-volume, instead of a typical sequence of 3D convolutions, drastically
augmenting the runtime speed of current models. In our experiments, we achieve real-time inference
speed, in the range of 5–32 ms, for 1216 368 input stereo images on the Jetson TX2, Jetson Xavier,
and Jetson Nano embedded devices.
|
Ricaurte P, Chilán C, Cristhian A. Aguilera, Boris X. Vintimilla, & Angel D. Sappa. (2014). Feature Point Descriptors: Infrared and Visible Spectra. Sensors Journal, Vol. 14, pp. 3690–3701.
Abstract: This manuscript evaluates the behavior of classical feature point descriptors when they are used in images from long-wave infrared spectral band and compare them with the results obtained in the visible spectrum. Robustness to changes in rotation, scaling, blur, and additive noise are analyzed using a state of the art framework. Experimental results using a cross-spectral outdoor image data set are presented and conclusions from these experiments are given.
|
Juan A. Carvajal, Dennis G. Romero, & Angel D. Sappa. (2016). Fine-tuning based deep covolutional networks for lepidopterous genus recognition. In XXI IberoAmerican Congress on Pattern Recognition (pp. 1–9).
Abstract: This paper describes an image classication approach ori- ented to identify specimens of lepidopterous insects recognized at Ecuado- rian ecological reserves. This work seeks to contribute to studies in the area of biology about genus of butter ies and also to facilitate the reg- istration of unrecognized specimens. The proposed approach is based on the ne-tuning of three widely used pre-trained Convolutional Neural Networks (CNNs). This strategy is intended to overcome the reduced number of labeled images. Experimental results with a dataset labeled by expert biologists, is presented|a recognition accuracy above 92% is reached. 1 Introductio
|
Juan A. Carvajal, Dennis G. Romero, & Angel D. Sappa. (2017). Fine-tuning deep convolutional networks for lepidopterous genus recognition. Lecture Notes in Computer Science, Vol. 10125 LNCS, pp. 467–475.
|
Jorge L. Charco, A. D. S., Boris X. Vintimilla, Henry O. Velesaca. (2022). Human Body Pose Estimation in Multi-view Environments. In ICT Applications for Smart Cities Part of the Intelligent Systems Reference Library book series (Vol. 224, pp. 79–99).
|
Jorge L. Charco, A. D. S., Boris X. Vintimilla. (2022). Human Pose Estimation through A Novel Multi-View Scheme. In Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications VISIGRAPP 2022 (Vol. 5, pp. 855–862).
Abstract: This paper presents a multi-view scheme to tackle the challenging problem of the self-occlusion in human
pose estimation problem. The proposed approach first obtains the human body joints of a set of images,
which are captured from different views at the same time. Then, it enhances the obtained joints by using a
multi-view scheme. Basically, the joints from a given view are used to enhance poorly estimated joints from
another view, especially intended to tackle the self occlusions cases. A network architecture initially proposed
for the monocular case is adapted to be used in the proposed multi-view scheme. Experimental results and
comparisons with the state-of-the-art approaches on Human3.6m dataset are presented showing improvements
in the accuracy of body joints estimations.
|
Angel D. Sappa. (2022). ICT Applications for Smart Cities. In Intelligent Systems Reference Library (Vol. 224).
|
Patricia L. Suarez, Angel D. Sappa, & Boris X. Vintimilla. (2019). Image patch similarity through a meta-learning metric based approach. In 15th International Conference on Signal Image Technology & Internet based Systems (SITIS 2019); Sorrento, Italia (pp. 511–517).
Abstract: Comparing images regions are one of the core methods used on computer vision for tasks like image classification, scene understanding, object detection and recognition. Hence, this paper proposes a novel approach to determine similarity of image regions (patches), in order to obtain the best representation of image patches. This problem has been studied by many researchers presenting different approaches, however, the ability to find the better criteria to measure the similarity on image regions are still a challenge. The present work tackles this problem using a few-shot metric based meta-learning framework able to compare image regions and determining a similarity measure to decide if there is similarity between the compared patches. Our model is training end-to-end from scratch. Experimental results
have shown that the proposed approach effectively estimates the similarity of the patches and, comparing it with the state of the art approaches, shows better results.
|
Patricia L. Suarez, Angel D. Sappa, Boris X. Vintimilla, & Riad I. Hammoud. (2019). Image Vegetation Index through a Cycle Generative Adversarial Network. In Conference on Computer Vision and Pattern Recognition Workshops (CVPR 2019); Long Beach, California, United States (pp. 1014–1021).
Abstract: This paper proposes a novel approach to estimate the
Normalized Difference Vegetation Index (NDVI) just from
an RGB image. The NDVI values are obtained by using
images from the visible spectral band together with a synthetic near infrared image obtained by a cycled GAN. The
cycled GAN network is able to obtain a NIR image from
a given gray scale image. It is trained by using unpaired
set of gray scale and NIR images by using a U-net architecture and a multiple loss function (gray scale images are
obtained from the provided RGB images). Then, the NIR
image estimated with the proposed cycle generative adversarial network is used to compute the NDVI index. Experimental results are provided showing the validity of the proposed approach. Additionally, comparisons with previous
approaches are also provided.
|
Xavier Soria, & Angel D. Sappa. (2018). Improving Edge Detection in RGB Images by Adding NIR Channel. In 14th IEEE International Conference on Signal Image Technology & Internet based Systems (SITIS 2018) (pp. 266–273).
|