Interactive Multi-Class Tiny-Object Detection

TL;DR

This paper introduces C3Det, an interactive multi-class tiny-object detection method that uses minimal user inputs to efficiently annotate tiny objects across multiple classes, outperforming existing approaches in accuracy and speed.

Contribution

The paper presents a novel interactive annotation framework for multi-class tiny objects, combining local and global context via late-fusion and feature correlation, with demonstrated efficiency improvements.

Findings

Outperforms existing methods in mAP with fewer clicks

Achieves 2.85x faster annotation speed in user studies

Yields significantly lower task load compared to manual annotation

Abstract

Annotating tens or hundreds of tiny objects in a given image is laborious yet crucial for a multitude of Computer Vision tasks. Such imagery typically contains objects from various categories, yet the multi-class interactive annotation setting for the detection task has thus far been unexplored. To address these needs, we propose a novel interactive annotation method for multiple instances of tiny objects from multiple classes, based on a few point-based user inputs. Our approach, C3Det, relates the full image context with annotator inputs in a local and global manner via late-fusion and feature-correlation, respectively. We perform experiments on the Tiny-DOTA and LCell datasets using both two-stage and one-stage object detection architectures to verify the efficacy of our approach. Our approach outperforms existing approaches in interactive annotation, achieving higher mAP with fewer clicks. Furthermore, we validate the annotation efficiency of our approach in a user study where it is shown to be 2.85x faster and yield only 0.36x task load (NASA-TLX, lower is better) compared to manual annotation. The code is available at https://github.com/ChungYi347/Interactive-Multi-Class-Tiny-Object-Detection.