| Presentation | 2015-02-24 The quantization noise reduction by blind deconvolution incorporating adaptive filter based on the structure image Takuya Miwa, Hironori Yamauchi, Tomonori Izumi, Kazunori Saito, Yohei Fukumizu, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In this paper, we propose a reduction method of quantization noise, which is based on blind deconvolution using total variation (TV). Quantization noises such as block noise and mosquito noise are caused by an image compression such as JPEG standard. Although noise reduction techniques against the significant deterioration of image quality, have been proposed the fine texture of the image included in high-frequency component inevitably suffers the blur accompanied by these noise reduction techniques. In order to suppress the blur on fine texture, we first separate the image into the structure image and the texture image by TV regularization. Then, an adaptive Gaussian filter is applied to the texture image where the intensity of the filter at each position is controlled according to the variance of the structure image. The adaptive filter is incorporated into the iteration of the blind deconvolution using TV. Our experimental result shows that the proposed method achieves higher signal-to-noise ratio (SNR) for significantly degraded images by quantization noise compared to the conventional method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | image restoration / quantization noise / total variation / blind deconvolution |
| Paper # | ITS2014-48,IE2014-75 |
| Date of Issue |
| Conference Information | |
| Committee | ITS |
|---|---|
| Conference Date | 2015/2/16(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 | Intelligent Transport Systems Technology (ITS) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | The quantization noise reduction by blind deconvolution incorporating adaptive filter based on the structure image |
| Sub Title (in English) | |
| Keyword(1) | image restoration |
| Keyword(2) | quantization noise |
| Keyword(3) | total variation |
| Keyword(4) | blind deconvolution |
| 1st Author's Name | Takuya Miwa |
| 1st Author's Affiliation | () |
| 2nd Author's Name | Hironori Yamauchi |
| 2nd Author's Affiliation | |
| 3rd Author's Name | Tomonori Izumi |
| 3rd Author's Affiliation | |
| 4th Author's Name | Kazunori Saito |
| 4th Author's Affiliation | |
| 5th Author's Name | Yohei Fukumizu |
| 5th Author's Affiliation | |
| Date | 2015-02-24 |
| Paper # | ITS2014-48,IE2014-75 |
| Volume (vol) | vol.114 |
| Number (no) | 459 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |