Accurate Metaphase Chromosome Classification Using Faster R-CNN and Deep LSTM Integration

The precise classification of metaphase chromosomes is essential for cytogenetic analysis, facilitating the detection of genetic diseases and anomalies. This paper introduces an improved methodology that combines Faster R-CNN (FR-CNN) for efficient feature extraction with Deep Long Short-Term Memory (LSTM) networks for sequence learning to categorize metaphase chromosomes. The Faster R-CNN model effectively identifies chromosome sections, but the Deep LSTM captures sequential dependencies within chromosome structures. This hybrid architecture improves classification accuracy while decreasing computational overhead relative to conventional techniques. Experimental results indicate that the suggested method attains an accuracy of 94%, a precision of 93%, a recall of 95%, and an F1-score of 94% on benchmark datasets, markedly surpassing current methodologies. This provides a dependable option for automated chromosomal analysis in clinical environments.