Jiangxi University of Finance and Economics

About: Jiangxi University of Finance and Economics is a education organization based out in Nanchang, China. It is known for research contribution in the topics: Fuzzy logic & China. The organization has 2865 authors who have published 3556 publications receiving 41567 citations.

Cryptanalysis of image scrambling based on chaotic sequences and Vigenère cipher

Abstract: Recently, an image scrambling scheme based on chaos theory and Vigenere cipher was proposed. The scrambling process is firstly to shift each pixel by sorting a chaotic sequence as Vigenere cipher, and then the pixel positions are shuffled by sorting another chaotic sequence. In this study, we analyze the security weakness of this scheme. By applying the combination of chosen-plaintext attack and differential attack, we propose two efficient cryptanalysis methods. Results show that all the keystream can be revealed. The original image scrambling scheme can be remedied by leveraging the MD5 hash value of the plain image as the initial condition of the chaotic system.

A Gated Recurrent Network With Dual Classification Assistance for Smoke Semantic Segmentation

Abstract: Smoke has semi-transparency property leading to highly complicated mixture of background and smoke. Sparse or small smoke is visually inconspicuous, and its boundary is often ambiguous. These reasons result in a very challenging task of separating smoke from a single image. To solve these problems, we propose a Classification-assisted Gated Recurrent Network (CGRNet) for smoke semantic segmentation. To discriminate smoke and smoke-like objects, we present a smoke segmentation strategy with dual classification assistance. Our classification module outputs two prediction probabilities for smoke. The first assistance is to use one probability to explicitly regulate the segmentation module for accuracy improvement by supervising a cross-entropy classification loss. The second one is to multiply the segmentation result by another probability for further refinement. This dual classification assistance greatly improves performance at image level. In the segmentation module, we design an Attention Convolutional GRU module (Att-ConvGRU) to learn the long-range context dependence of features. To perceive small or inconspicuous smoke, we design a Multi-scale Context Contrasted Local Feature structure (MCCL) and a Dense Pyramid Pooling Module (DPPM) for improving the representation ability of our network. Extensive experiments validate that our method significantly outperforms existing state-of-art algorithms on smoke datasets, and also obtain satisfactory results on challenging images with inconspicuous smoke and smoke-like objects.