Embedded Files
Cauchy aperture: Camera aperture design for extending depth-of-field
Cauchy aperture: Camera aperture design for extending depth-of-field
A pupil function of aperture in image capturing systems is theoretically derived such that one can perfectly reconstruct an all-in-focus image through linear filtering of the focal stack. The perfect reconstruction filters are also designed based on the derived pupil function. The designed filters are space-invariant; hence the presented method does not require region segmentation.
Akira Kubota, Kazuya Kodama, Asami Ito, "Cauchy Aperture and Perfect Reconstruction Filters for Extending Depth-of-Field from Focal Stack," IEICE TRANSACTIONS on Information and Systems, Vol.E102-D, No.11, pp.2093-2100, Nov. 2019.
Akira Kubota, Synthesis filter bank and pupil function for perfect reconstruction of all-in-focus image from focal stack, SPIE International Conference on Quality Control by Artificial Vision (QCAV) 2017, Vol.103380A, doi:10.1117/12.2266960, May, 2017.
Audio feature using sub-band correlations for music genre classification
Audio feature using sub-band correlations for music genre classification
We present novel low-level audio features based on sub-band correlations for effective music genre classification. Under the assumption that SVM is used for classifier learning, the experimental results on GTZAN data set showed that the proposed method demonstrated the best accuracy of 81.5%, outperforming the conventional methods. (Currently, we achieved 87.1% by including the conventional audio features such as MFCC)
Takuya Kobayashi, Akira Kubota, Yusuke Suzuki, "Audio Feature Extraction Based on Sub-Band Signal Correlations for Music Genre Classification," 2018 IEEE International Symposium on Multimedia (ISM), Taichung, 2018, pp. 180-181.
Mixed-Reality System for Self-Study Spaces Improving Concentration
Mixed-Reality System for Self-Study Spaces Improving Concentration
In recent years, demands for self-study spaces have increased among students and office workers. This project develops a novel system that creates mixed-reality spaces to improve concentration using a head-mounted display (HMD). The system automatically obscures visual areas other than ones the users need to concentrate on by rendering virtual occluding objects through HMD over real captured scenes.
Kanta Michioka and Akira Kubota, "Development of an Mixed-Reality System for Self-Study Spaces Improving Concentration," The Tenth International Workshop on Image Media Quality and its Applications (IMQA 2020), to appear
Focal stack refocusing through linear filtering
Focal stack refocusing through linear filtering
This project gives a digital refocusing method that transforms the captured focal stack directly into a new focal stack under different focus settings. Assuming Lambertian scenes with no occlusions, this paper theoretically shows that there exist a set of filters that perfectly reconstructs the focal stack under Gaussian aperture from that captured under Cauchy one. The perfect reconstruction filters are derived in linear and space-invariant using a layered scene representation. Numerical simulations using synthetic focal stacks showed that the root-mean- squared errors are quite small and less than 1.0E-9, indicating the derived filters allow perfect reconstruction.
Asami Ito, Akira Kubota, Kazuya Kodama, "Deriving Perfect Reconstruction Filter Bank for Focal Stack Refocusing," Asian Conference on Pattern Recognition (ACPR) 2019, paper#114, pp.1-10, Nov. 2019.
Other current/past research projects will be listed soon.
Page updated
Report abuse