“In the field of end-to-end object detection, a novel query-selection encoder (QSE) has been introduced to enhance training convergence speed and detection accuracy. Expert researchers incorporated a hierarchical feature-aware attention (HFA) mechanism, which effectively suppresses similar feature representations and highlights discriminative ones, accelerating the feature selection process. This method is highly versatile in accommodating both CNN- and Transformer-based detectors, laying a foundation for the construction of more efficient detection systems.”