TOD-CNN: An Effective Convolutional Neural Network for Tiny Object Detection in Sperm Videos
This work addresses a domain-specific challenge in medical imaging for sperm quality analysis, showing incremental improvements in detection accuracy.
The paper tackles the problem of detecting tiny objects like sperm in microscopic videos, achieving 85.60% AP50 accuracy for real-time detection using a dataset of 111 videos with over 278,000 annotated objects.
The detection of tiny objects in microscopic videos is a problematic point, especially in large-scale experiments. For tiny objects (such as sperms) in microscopic videos, current detection methods face challenges in fuzzy, irregular, and precise positioning of objects. In contrast, we present a convolutional neural network for tiny object detection (TOD-CNN) with an underlying data set of high-quality sperm microscopic videos (111 videos, $>$ 278,000 annotated objects), and a graphical user interface (GUI) is designed to employ and test the proposed model effectively. TOD-CNN is highly accurate, achieving $85.60\%$ AP$_{50}$ in the task of real-time sperm detection in microscopic videos. To demonstrate the importance of sperm detection technology in sperm quality analysis, we carry out relevant sperm quality evaluation metrics and compare them with the diagnosis results from medical doctors.