author: qingshan liu

... Highlight removal from a single input image has been studied for decades. In this section, we briefly review ... remove the specular highlight by defining a new L2 chromaticity. Yang et al. [10] propose a fast real-time implementation by ...

... diet analysis and food recommendation. Due to the similar appearance and shape between different foods, it is quite ... personal diet using the easily accessible food images, which could be captured by various smart devices, such as ...

... TF2 TF3 RF1 RF2 RF3 Mutil - Scale Fusion FCM Preliminary result RRF1 RRF2 ... TF2 , TF3 ) and completed 3D RGB features ( RRF1 ) with the proposed 3D RGB ... Hsu et al . [ 11 ] motivate using pairwise constraints to learn clustering ...

22nd International Conference, ICONIP 2015, Istanbul, Turkey, November 9-12, 2015 ... hand, right hand, foot and tongue motor imagery tasks. For the purpose of ... feet. 5 subjects had previously performed BCI experiments and 9 subjects ...

... 3D central differ- ence separable residual long and short - term memory ( LSTM ) coordinate attention ( 3D CRLCA ) , which addresses these issues . Our network uti- lizes 3D central difference separable convolution ( 3D CDSC ) to ...

... bookfi. org/book/805420. (in Russian) 4. Good, I.J.: The interaction algorithm and practical fourier analysis. J. R. Statistic. Soc. Ser. B 20, 361–372 (1958) 5. Dorogov, A.Y., Shestopalov, M.Y.: Neirosetevoe Modelirovanie Regulyarnyh ...

... UCI repository1. Their statistical characteristics are given in Table1 and below a brief verbal description of each data set is given: – ECOLI, the UCI Ecoli data set, containing protein localization sites, originally with 8 attributes ...

... This phenomenon stems from the fact that the primary functions of the SVE and PVE modules are geared towards ... information transfer process of the optical. 4 Conclusion The scarcity of training samples , the influence of irrelevant ...

... ezviz fisheye camera to record content from Problem-Based Learning classes and meetings [1] at a medical college in China. A total of 3,573 frames and 31,915 samples were selected from a large number of videos to form the dataset. Eight ...