| Presentation | 2014-09-18 Architecture Development for the Real-time Computer-Aided Diagnosis of Colorectal Endoscopic Images with NBI Magnification Tetsushi KOIDE, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | With the increase in patients of colorectal cancer in recent years, the requirement of the effective diagnostic support system is increased. In this study, we propose an architecture for real-time computer-aided diagnosis (CAD) system of colorectal endoscopic images with narrow-band imaging (NBI) magnification. The proposed algorithm consists of three main processings: (1) the feature extraction which is based on a bag-of-features (BoF) representation of local features and the Dense Scale-Invariant Feature Transform (D-SIFT), (2) the feature transformation based on the visual word histogram construction with k-means clustering, and (3) the type classification based on the support vector machine. We develop the hardware-oriented algorithm for each processing and investigate a trade-off of the recognition precision of these algorithms and the speedup by the FPGA implementation for the real time processing of full high definition (1920 x 1080 pixels) image. In this paper, we show the hardware algorithms, simulation results, architecture designs, and these FPGA implementations. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Colorectal Endoscopic Images / Computer-Aided Diagnosis (CAD) / Real-Time Image Processing / Bag-of-Features (BoF) / Dense Scale-Invariant Feature Transform (D-SIFT) / k-means Clustering / Support Vector Machine (SVM) / FPGA |
| Paper # | RECONF2014-17 |
| Date of Issue |
| Conference Information | |
| Committee | RECONF |
|---|---|
| Conference Date | 2014/9/11(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Reconfigurable Systems (RECONF) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Architecture Development for the Real-time Computer-Aided Diagnosis of Colorectal Endoscopic Images with NBI Magnification |
| Sub Title (in English) | |
| Keyword(1) | Colorectal Endoscopic Images |
| Keyword(2) | Computer-Aided Diagnosis (CAD) |
| Keyword(3) | Real-Time Image Processing |
| Keyword(4) | Bag-of-Features (BoF) |
| Keyword(5) | Dense Scale-Invariant Feature Transform (D-SIFT) |
| Keyword(6) | k-means Clustering |
| Keyword(7) | Support Vector Machine (SVM) |
| Keyword(8) | FPGA |
| 1st Author's Name | Tetsushi KOIDE |
| 1st Author's Affiliation | Research Institute for Nanodevice and Bio System (NBS):Graduate School of Advanced Sciences of Matter, Hiroshima University() |
| Date | 2014-09-18 |
| Paper # | RECONF2014-17 |
| Volume (vol) | vol.114 |
| Number (no) | 223 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |