Conv-MTD: A CNN Based Multi-Label Medical Tubes Detection and Classification Model to Facilitate Resource-constrained Point-of-care Devices.

Computer-aided detection through deep learning is becoming a prevalent approach across various fields, including detection of anomalies in medical procedures. One such medical procedure involves the placement of medical tubes to provide nutrition or other medical procedures in critically ill patients. Medical tube placement can be highly complex and prone to subjective errors.