Automated Palynological Analysis System: Integrating Deep Metric Learning and $U^{2}$-Net Detection in $H\infty$ bright field microscopy

TL;DR

This paper introduces an automated microscopy system that combines advanced deep learning techniques with robust mechanical control to rapidly and accurately analyze pollen grains, significantly reducing analysis time.

Contribution

The system uniquely integrates $U^{2}$-Net detection with DINOv2 Vision Transformer and Gradient-Weighted Attention for efficient, interpretable pollen analysis, advancing automation in melissopalynology.

Findings

Achieved 95.8% classification recall.

Realized a 6x speedup over manual analysis.

Enabled precise counting, classification, and morphological analysis.

Abstract

Traditional melissopalynology is a time-consuming and subjective process, often taking 4-6 hours per sample. We present an automated, high-throughput microscopy system that integrates $H\infty$ robust mechanical control with advanced deep learning pipelines for the precise counting, classification, and morphological analysis of pollen grains from Bio Bio region in south central territory in Chile. Our system employs $U^{2}$-Net for salient object detection and a DINOv2 Vision Transformer backbone trained via Deep Metric Learning for classification. By integrating Gradient-Weighted Attention, the model provides human-interpretable texture and diagnostic feature annotations. The system achieves a 95.8$\%$ classification recall and a 6x processing speedup compared to manual expert analysis.