1st Tiny Object Detection (TOD) Challenge

Real-world Recognition from Low-quality Inputs (RLQ)

August 23rd - 28th, SEC, GLASGOW

Introduction

How is the robustness of the current state-of-the-art for recognition and detection algorithms in non-ideal visual environments? While the visual recognition research has made tremendous progress in recent years, most models are trained, applied, and evaluated on high-quality (HQ) visual data. However, in many emerging applications such as robotics and autonomous driving, the performances of visual sensing and analytics are largely jeopardized by low-quality(LQ) visual data acquired from unconstrained environments, suffering from various types of degradation such as low resolution, noise, occlusion, motion blur, contrast, brightness, sharpness, out-of-focus etc. We are proposing to organise the 2nd RLQ workshop in conjunction with ECCV 2020 to provide an integrated forum for both low-level and high-level vision researchers to review the recent progress of robust recognition models from LQ visual data and the novel image restoration algorithms.

[Challenge 1] In conjunction with the workshop, we will hold the 1st Tiny Object Detection (TOD) Challenge. This challenge targets at establishing a baseline for tiny person detection by presenting a new benchmark and various approaches, opening up a promising direction for tiny object detection in the wild. The new benchmark, named TinyPerson, spans challenges including extreme low-resolution, background diversity, multi-objects, part-invisibility, and various complex backgrounds that are far beyond those in existing datasets.

[Challenge 2] We will hold the first image restoration challenge on Under-Display Camera (UDC). The new trend of full-screen devices encourages us to position a camera behind a screen. Removing the bezel and centralizing the camera under the screen brings larger display-to-body ratio and enhances eye contact in video chat, but also causes image degradation. we focus on a newly-defined Under-Display Camera (UDC), as a novel real-world single image restoration problem. We will release the UDC dataset for training and testing, and rank the algorithms according to both performance and efficiency.

Important Dates

Description Date
Full Paper Submission Deadline 25 June 2020 (23:59 UTC-0)
Notification to Authors 10 July 2020 (23:59 UTC-0)
Camera-Ready Deadline 15 July 2020 (23:59 UTC-0)
Challenge Paper Submission Deadline 1 June 2020 (23:59 UTC-0)
Challenge Paper Notification 10 June 2020 (23:59 UTC-0)
Challenge Paper Camera-Ready 15 June 2020 (23:59 UTC-0)
Workshop Presentation 23 August 2020 (Whole Day)