paper-with-me

Papers

Matching Targets Across Domains with RADON, the Re-Identification Across Domain Network

2021-05-25 · Cassandra Burgess, Cordelia Neisinger, Rafael Dinner

We present a novel convolutional neural network that learns to match images of an object taken from different viewpoints or by different optical sensors. Our Re-Identification Across Domain Network (RADON) scores pairs of input images from different domains on similarity. Our approach extends previous work on Siamese networks and modifies them to more challenging use cases, including low- and no-shot learning, in which few images of a specific target are available for training. RADON shows strong performance on cross-view vehicle matching and cross-domain person identification in a no-shot learning environment.

📄 PDF Abstract BibTeX arXiv:2105.12056

Code (0)

등록된 구현이 없습니다.

Tasks

Person Identification

Similar Papers 제목 키워드 기반

A Radon-Nikodým Perspective on Anomaly Detection: Theory and Implications

2025-02-25 · Shlok Mehendale, Aditya Challa, Rahul Yedida, Sravan Danda 외

Which principle underpins the design of an effective anomaly detection loss function? The answer lies in the concept of Radon-Nikod\'ym theorem, a fundamental concept in measure theory. The key insight from this article …

Anomaly DetectionSupervised Anomaly DetectionTime SeriesUnsupervised Anomaly Detection

Detection of Line Artefacts in Lung Ultrasound Images of COVID-19 Patients via Non-Convex Regularization

2020-05-06 · Oktay Karakuş, Nantheera Anantrasirichai, Amazigh Aguersif, Stein Silva 외

In this paper, we present a novel method for line artefacts quantification in lung ultrasound (LUS) images of COVID-19 patients. We formulate this as a non-convex regularisation problem involving a sparsity-enforcing, Ca…

Generating Binary Tags for Fast Medical Image Retrieval Based on Convolutional Nets and Radon Transform

2016-04-16 · Xinran Liu, Hamid. R. Tizhoosh, Jonathan Kofman

Content-based image retrieval (CBIR) in large medical image archives is a challenging and necessary task. Generally, different feature extraction methods are used to assign expressive and invariant features to each image…

Content-Based Image RetrievalGeneral ClassificationImage RetrievalMedical Image Retrieval+1

Hierarchical Hybrid Sliced Wasserstein: A Scalable Metric for Heterogeneous Joint Distributions

2024-04-23 · Khai Nguyen, Nhat Ho

Sliced Wasserstein (SW) and Generalized Sliced Wasserstein (GSW) have been widely used in applications due to their computational and statistical scalability. However, the SW and the GSW are only defined between distribu…

Inverse Airborne Optical Sectioning

2022-07-27 · Rakesh John Amala Arokia Nathan, Indrajit Kurmi, Oliver Bimber

We present Inverse Airborne Optical Sectioning (IAOS) an optical analogy to Inverse Synthetic Aperture Radar (ISAR). Moving targets, such as walking people, that are heavily occluded by vegetation can be made visible and…