The paper introduces DAVE, a novel low-shot counting and detection method that combines the strengths of density-based and detection-based approaches. DAVE addresses the limitations of existing methods by first generating a high-recall detection set and then verifying the detections to identify and remove outliers, thereby improving both recall and precision. This approach enhances the accuracy of total counts and detection quality. DAVE outperforms state-of-the-art density-based counters by approximately 20% in total count Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), and outperforms detection-based counters by a similar margin in detection metrics. It also sets a new state-of-the-art in zero-shot and text-prompt-based counting. The method is evaluated on various benchmarks, demonstrating its effectiveness in challenging scenarios. The code and models are available on GitHub.The paper introduces DAVE, a novel low-shot counting and detection method that combines the strengths of density-based and detection-based approaches. DAVE addresses the limitations of existing methods by first generating a high-recall detection set and then verifying the detections to identify and remove outliers, thereby improving both recall and precision. This approach enhances the accuracy of total counts and detection quality. DAVE outperforms state-of-the-art density-based counters by approximately 20% in total count Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), and outperforms detection-based counters by a similar margin in detection metrics. It also sets a new state-of-the-art in zero-shot and text-prompt-based counting. The method is evaluated on various benchmarks, demonstrating its effectiveness in challenging scenarios. The code and models are available on GitHub.