Dmytro Mishkin

О себе

Work in computer vision since 2012.

Commercial research: developed core technology for — personalized, visual based shopping app at Clear Research.

Academy research: developed MODS — state-of-art algorithm for wide baseline stereo matching and LSUV — state-of-art method for convolution neural networks initialization.

Co-founder of Szkocka Research Group — Ukrainian open research community for computer science.

Kaggler, 9th place out of 1049 teams at Kaggle National Data Science Bowl.

Open source contributor: deep learning libraries — caffe and keras.
Reviewer for IEEE Transactions on Pattern Analysis and Machine Intelligence.