Systems and methods for detecting regions of interests (ROIs) in biomedical images are described herein. A computing system may identify a biomedical image having an ROI. The computer system may apply an object detection model to the biomedical image. The object detection model may generate a feature map using the biomedical image. The object detection model may generate an anchor box corresponding to a portion of the pixels in the feature map. The computing system may apply an instance segmentation model to identify a segment of the biomedical image within the anchor box corresponding to the ROI. The computer system may provide an output based on the segment to identify the ROI in the biomedical image.