Application-specific processor for local-binary-patterns generation

Early detection of polyps could result in better prevention against colorectal cancer (CRC). The classical method adopted for this purpose is the clinical inspection. However, this type of detection suffer from several limitations such as the medical level of expertise and the nature and appearance of the polyp itself. Therefore, there is an essential need for support systems which help in the detection process. In this work, Local-Binary-Pattern (LBP) is utilized as a feature extractor for the purpose of real-time detection of polyps. LBP is known for its rich texture analysis and simple computation of the detection tasks. In order to take into account the low power requirement of LBP, an Application-Specific Integrated Circuits (ASIC) flow is implemented which turned out to be essential for the considered detection. Synthesis results demonstrate that the proposed detection method achieves a considerable reduction in the power consumption as compared to other literature works. Additionally, the results confirm that the proposed method is capable of running under ultra-low power requirements.