Overview
A study on the use of the Faster RCNN R101 R101 FPN model within the Facebook detectron2 model zoo on the analysis of the detection of multiple levels of burn injuries within the same image with the use of an augmented dataset for training.
Skin Burn Detection
A study on the use of the Faster RCNN R101 R101 FPN model within the Facebook detectron2 model zoo on the analysis of the detection of multiple levels of burn injuries within the same image with the use of an augmented dataset for training.
Last updated December 13, 2022
Tech Stack
Python
Primary language: Python
Engineering Decisions
- Prioritize maintainability and clear interfaces for long-term evolution.
- Design around observability and operational correctness from early stages.
- Balance product velocity with platform reliability tradeoffs.
Outcomes and Future Direction
- Demonstrates end-to-end ownership across architecture and product execution.
- Shows practical capability in shipping production-style systems.
- Future work includes deeper benchmarks, testing, and system hardening.