Benchmarking Deep Learning-Based Object Detection Models on Feature Deficient Astrophotography Imagery Dataset

TL;DR

This paper benchmarks deep learning object detection models on a novel astrophotography dataset with sparse night-sky images, revealing challenges in feature-deficient conditions and highlighting the need for specialized models.

Contribution

It introduces MobilTelesco, a new astrophotography dataset, and evaluates existing models' performance on sparse, feature-deficient night-sky images.

Findings

Detection models struggle with feature-deficient astrophotography images.

Benchmark results highlight the need for specialized detection approaches.

Challenges in applying standard models to non-commercial, sparse imagery.

Abstract

Object detection models are typically trained on datasets like ImageNet, COCO, and PASCAL VOC, which focus on everyday objects. However, these lack signal sparsity found in non-commercial domains. MobilTelesco, a smartphone-based astrophotography dataset, addresses this by providing sparse night-sky images. We benchmark several detection models on it, highlighting challenges under feature-deficient conditions.