Fabian Fallas-Moya
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Fabian Fallas-Moya has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Danny Xie-Li
Instituto Tecnológico de Costa Rica
2 shared publications - Amir Sadovnik
Oak Ridge National Laboratory
1 shared publication B. Meden
Faculty of Computer & Information Science, University of Ljubljana
1 shared publicationFrancisco Torres-Rojas
Instituto Tecnológico de Costa Rica
1 shared publicationH. Stark
Departments of Medical Imaging and Statistical Sciences, University of Toronto
1 shared publication- Hairong Qi
Department of Electrical Engineering and Computer Science, University of Tennessee, Knoxville, TN 37996, USA
1 shared publication
Latest Publications
- Open Access
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1457-1464
Peeling Back the Pixels: Stabilizing Latent Diffusion for Data-Efficient Generation of Pineapple Fields
While Latent Diffusion Models (LDMs) excel at synthetic data generation, they suffer from severe training instability on the small, homogeneous UAV datasets typical of precision agriculture. In this work, we investigate these architectural bottlenecks by training a Denoising Diffusion Probabilistic Model (DDPM) from scratch on constrained pineapple plantation imagery. Our systematic comparison reveals that continuous latent spaces (β-VAE) preserve complex crop structures, whereas discrete architectures (VQ-VAE) fail under scarce data conditions. Additionally, we demonstrate that replacing destabilizing global attention mechanisms with a purely convolutional U-Net backbone ensures smooth convergence in repetitive visual domains. Ultimately, combining this robust architecture with DDIM sampling yields a stable, highly efficient framework for low-data agricultural image synthesis. Code and models will be available at https://github.com/imagine-laboratory/Latent-Diffusion.
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1457-1464
Open AccessPeeling Back the Pixels: Stabilizing Latent Diffusion for Data-Efficient Generation of Pineapple Fields
While Latent Diffusion Models (LDMs) excel at synthetic data generation, they suffer from severe training instability on the small, homogeneous UAV datasets typical of precision agriculture. In this work, we investigate these architectural bottlenecks by training a Denoising Diffusion Probabilistic Model (DDPM) from scratch on constrained pineapple plantation imagery. Our systematic comparison reveals that continuous latent spaces (β-VAE) preserve complex crop structures, whereas discrete architectures (VQ-VAE) fail under scarce data conditions. Additionally, we demonstrate that replacing destabilizing global attention mechanisms with a purely convolutional U-Net backbone ensures smooth convergence in repetitive visual domains. Ultimately, combining this robust architecture with DDIM sampling yields a stable, highly efficient framework for low-data agricultural image synthesis. Code and models will be available at https://github.com/imagine-laboratory/Latent-Diffusion.
- Open Access
- 8 April 2025
- Published by Instituto Tecnologico de Costa Rica in Revista Tecnología en Marcha
- p. ág. 100-111
Adding a teaching “assistant”: improving the quality of pseudo-labels for semi-supervised object detection
This paper focuses on semi-supervised object detection (SS-OD) for its tolerance to small amounts of training samples, which is common in real-world applications. Pseudo-label-based approaches have been the mainstream for SS-OD. In this paper, we first show the impact of accurate pseudo-labeling and the challenge of producing such labels. In contrast to prior research that predominantly focused on refining the main model to enhance localization, this paper introduces a novel strategy, where a standalone “Teaching Assistant” or simply “Assistant” is involved in the popular Teacher/Student paradigm to improve the quality of pseudo-labels. This “Assistant” can be plugged into any existing Teacher/Student-based framework without having to fine-tune the original Teacher/Student model. We exploit two “Assistant” models, both of which center around the non-maximum suppression (NMS) method -- a popular technique used to select only the promising bounding boxes. The first “Assistant” model is referred to as the “pre-NMS” assistant that refines the candidate bounding box scores for a better set of inputs to the NMS process. The second “Assistant” model is referred to as the “post-NMS” assistant which takes advantage of SOTA segmentation models to improve the output from the NMS process. We thoroughly evaluate the performance of pre-NMS vs. post-NMS and the impact of improved pseudo-labels on the OD performance. Experimental results on the COCO dataset demonstrate that post-NMS is better than SOTA methods.
- 8 April 2025
- Published by Instituto Tecnologico de Costa Rica in Revista Tecnología en Marcha
- p. ág. 100-111
Open AccessAdding a teaching “assistant”: improving the quality of pseudo-labels for semi-supervised object detection
This paper focuses on semi-supervised object detection (SS-OD) for its tolerance to small amounts of training samples, which is common in real-world applications. Pseudo-label-based approaches have been the mainstream for SS-OD. In this paper, we first show the impact of accurate pseudo-labeling and the challenge of producing such labels. In contrast to prior research that predominantly focused on refining the main model to enhance localization, this paper introduces a novel strategy, where a standalone “Teaching Assistant” or simply “Assistant” is involved in the popular Teacher/Student paradigm to improve the quality of pseudo-labels. This “Assistant” can be plugged into any existing Teacher/Student-based framework without having to fine-tune the original Teacher/Student model. We exploit two “Assistant” models, both of which center around the non-maximum suppression (NMS) method -- a popular technique used to select only the promising bounding boxes. The first “Assistant” model is referred to as the “pre-NMS” assistant that refines the candidate bounding box scores for a better set of inputs to the NMS process. The second “Assistant” model is referred to as the “post-NMS” assistant which takes advantage of SOTA segmentation models to improve the output from the NMS process. We thoroughly evaluate the performance of pre-NMS vs. post-NMS and the impact of improved pseudo-labels on the OD performance. Experimental results on the COCO dataset demonstrate that post-NMS is better than SOTA methods.
- 1 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1752-1756
Assessing the Impact of the Deceived Non Local Means Filter as a Preprocessing Stage in a Convolutional Neural Network Based Approach for Age Estimation Using Digital Hand X-Ray Images
In this work we analyze the impact of denoising, contrast and edge enhancement using the Deceived Non Local Means (DNLM) filter in a Convolutional Neural Network (CNN) based approach for age estimation using digital X-ray images from hands. The DNLM filter presents two parameters which control edge enhancement and denoising. Increasing levels were tested to assess the impact of both contrast enhancement and denoising in the CNN based model regression accuracy. Results obtained showed that contrast enhancement was important for preprocessing in a CNN based approach, given a statistically significant 42% lower root mean squared error, with comparable to previous state of the art results, using larger publicly available dataset. The obtained results suggest that both image enhancement and denoising can significantly improve results in a CNN based model.
- 1 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1752-1756
Assessing the Impact of the Deceived Non Local Means Filter as a Preprocessing Stage in a Convolutional Neural Network Based Approach for Age Estimation Using Digital Hand X-Ray Images
In this work we analyze the impact of denoising, contrast and edge enhancement using the Deceived Non Local Means (DNLM) filter in a Convolutional Neural Network (CNN) based approach for age estimation using digital X-ray images from hands. The DNLM filter presents two parameters which control edge enhancement and denoising. Increasing levels were tested to assess the impact of both contrast enhancement and denoising in the CNN based model regression accuracy. Results obtained showed that contrast enhancement was important for preprocessing in a CNN based approach, given a statistically significant 42% lower root mean squared error, with comparable to previous state of the art results, using larger publicly available dataset. The obtained results suggest that both image enhancement and denoising can significantly improve results in a CNN based model.
- 17 February 2018
- Published by Springer Nature
- p. 1-14
Object Recognition Using Hierarchical Temporal Memory
- 17 February 2018
- Published by Springer Nature
- p. 1-14
Object Recognition Using Hierarchical Temporal Memory