Tag: #Uncertain/Imprecise Labels

MATLAB implementation of Multi-target Extended Functions of Multiple Instances has been made public! It is available in our GitHub repository MT_eFUMI  MT_eFUMI is capable of learning multiple target spectral signatures from weakly- and inaccurately-labeled hyperspectral imagery. It is a generalization of the Function of Multiple Instances approach (FUMI). Additional details can be found in the […]

Congratulations to our labmates and collaborators: Dylan Stewart, Alina Zare, Sergio Marconi, Ben Weinstein, Ethan White, Sarah Grave, Stephanie Bohlman and Aditya Singh! Their paper, “RANDCROWNS: A Quantitative Metric for Imprecisely Labeled Tree Crown Delineation”, was recently accepted to IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021. In the paper, […]

Abstract: Supervised methods for object delineation in remote sensing require labeled ground-truth data. Gathering sufficient high quality ground-truth data is difficult, especially when targets are of irregular shape or difficult to distinguish from background or neighboring objects. Tree crown delineation provides key information from remote sensing images for forestry, ecology, and management. However, tree crowns […]

Abstract: Imprecise labels or label uncertainty are common problems in many real supervised and semi-supervised learning problems. However, most of the state-of-the-art supervised learning methods in the literature rely on accurate labels. Accurate labels are often either expensive, time-consuming, or even impossible to obtain in many real applications. There are many approaches in the literature […]

Congratulations to our labmates: Dylan Stewart, Anna Hampton, Alina Zare, Jeff Dale and James Keller!  Their paper, “The Weakly-Labeled Rand Index” was recently accepted to the IEEE International Geoscience and Remote Sensing Symposium (IGARSS). In their paper, the authors introduce an approach to quantify superpixel segmentation performance. Whereas traditional evaluation approaches require crisp segmentations, the […]

Abstract: Synthetic Aperture Sonar (SAS) surveys produce imagery with large regions of transition between seabed types. Due to these regions, it is difficult to label and segment the imagery and, furthermore, challenging to score the image segmentations appropriately. While there are many approaches to quantify performance in standard crisp segmentation schemes, drawing hard boundaries in […]

Congratulations to our labmates and collaborators Guohao Yu, Alina Zare, Hudanyun Sheng, Roser Matamala, Joel Reyes-Cabrera, Felix Fritschi and Thomas Juenger! Their paper, “Root Identification in Minirhizotron Imagery with Multiple Instance Learning”, was recently published in Machine Vision and Applications! Their paper explores the use of multiple instance learning to segment minirhizotron images of plant […]

Congratulations to our labmates Guohao Yu, Alina Zare and Hudanyun Sheng, as well as collaborators, Roser Matamala, Joel Reyes-Cabrera, Felix Fritschi and Thomas Juenger! Their paper, “Root Identification in Minirhizotron Imagery with Multiple Instance Learning”, was recently accepted to Machine Vision and Applications! Their paper explores the use of multiple instance learning to segment minirhizotron […]

Congratulations to GatorSense alumna, Xiaoxiao Du!  Her paper, titled “Multi-resolution Multi-modal Sensor Fusion For Remote Sensing Data with Label Uncertainty”, was recently published in IEEE Transactions on Geoscience and Remote Sensing.   Check it out here!

Abstract: In remote sensing, it is often challenging to acquire or collect a large dataset that is accurately labeled. This difficulty is usually due to several issues, including but not limited to the study site’s spatial area and accessibility, errors in the global positioning system (GPS), and mixed pixels caused by an image’s spatial resolution. […]