Measuring the respiratory rate of inpatients is important for detecting physical changes and signs of abnormality in medical field. Therefore, respiratory rate of inpatients is regularly measured. However, measuring respiratory rates places a heavy burden of healthcare workers. The reason is that respiratory rate is measured by visual observation. To reduce the burden of healthcare workers, we investigate a simple, non-invasive and automatic method of measuring respiratory rate. We focus on object detection algorithms. Object detection algorithms detect objects in the image. Then, bounding boxes are output at the locations of detected objects. We propose a method for measuring respiratory rate using the width of the bounding box.