no code implementations • 21 Jan 2019 • Sunder Ali Khowaja, Seok-Lyong Lee
We also propose the sequential combination of Inception-ResNetv2 and long-short-term memory network (LSTM) to leverage the temporal variances for improved recognition performance.