No abstract

No abstract

No abstract

Time series medical images are an important type of medical data that contain rich spatio-temporal information. Recently, computer-aided diagnosis algorithms are often used on honest-but-curious servers to analyze medical images, which introduces severe privacy concerns. This paper proposes a convolutional-LSTM network (HE-CLSTM) for analyzing encrypted medical image sequences. Specifically, several convolutional blocks and modified LSTM layers (HE-LSTM) are constructed to extract discriminative spatio-temporal features. Moreover, a weighted unit and sequence voting layer are designed to improve accuracy while reducing the missed diagnosis rate. The experimental results on two benchmarks prove the superiority of HE-CLSTM on analyzing encrypted medical image sequences.

No abstract