# On Utilizing Relationships for Transferable Few-Shot Fine-Grained Object Detection Source: [On Utilizing Relationships for Transferable Few-Shot Fine-Grained Object Detection]() Collection index: [llms.txt]() Data updated: 2025-10-28T13:56:56.488Z Authors: Ambar Pal, Arnau Ramisa, Amit Kumar K C, Ren\\'e Vidal Published: 2022-12-01T18:54:07.000Z ## Abstract State-of-the-art object detectors are fast and accurate, but they require a large amount of well annotated training data to obtain good performance. However, obtaining a large amount of training annotations specific to a particular task, i.e., fine-grained annotations, is costly in practice. In contrast, obtaining common-sense relationships from text, e.g., "a table-lamp is a lamp that sits on top of a table", is much easier. Additionally, common-sense relationships like "on-top-of" are easy to annotate in a task-agnostic fashion. In this paper, we propose a probabilistic model that uses such relational knowledge to transform an off-the-shelf detector of coarse object categories (e.g., "table", "lamp") into a detector of fine-grained categories (e.g., "table-lamp"). We demonstrate that our method, RelDetect, achieves performance competitive to finetuning based state-of-the-art object detector baselines when an extremely low amount of fine-grained annotations is available ($0.2\\%$ of entire dataset). We also demonstrate that RelDetect is able to utilize the inherent transferability of relationship information to obtain a better performance ($+5$ mAP points) than the above baselines on an unseen dataset (zero-shot transfer). In summary, we demonstrate the power of using relationships for object detection on datasets where fine-grained object categories can be linked to coarse-grained categories via suitable relationships. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. ## Original sources and implementation [Original paper]() [PDF]()