In order to solve the problem of insufficient available training data in the classification task of breast mass and calcification, a multi-view model based on secondary transfer learning was proposed combining with imaging characteristics of mammogram. Firstly, CBIS-DDSM (Curated Breast Imaging Subset of Digital Database for Screening Mammography) was used to construct the breast local tissue section dataset for the pre-training of the backbone network, and the domain adaptation learning of the backbone network was completed, so the backbone network had the essential ability of capturing pathological features. Then, the backbone network was secondarily transferred to the multi-view model and was fine-tuned based on the dataset of Mianyang Central Hospital. At the same time, the number of positive samples in the training was increased by CBIS-DDSM to improve the generalization ability of the network. The experimental results show that the domain adaption learning and data augmentation strategy improves the performance criteria by 17% averagely and achieves 94% and 90% AUC (Area Under Curve) values for mass and calcification respectively.

To facilitate the quick detection of primary user and discover the available channel in a wireless cognitive network, a novel spectrum sensing mechanism was proposed. The mechanism conducted spectrum sensing simultaneously with data transmission to reduce the sense time, and exploited the interference cancellation technique to cancel out the self-interference. Furthermore, by combining the Fourier transform results of multiple signals with different sampling rates, it could quickly detect the behavior of the wideband channel. Finally, during the transmission, the primary user was detected in time by the well-known signature matching method. The experimental results show that, the sense time is reduced by 50%, and the throughput gain is as large as 100%. In summary, the proposed scheme can effectively reduce the overhead of spectrum sensing and improve the efficiency of cognitive communication.

In this paper, two discrete Harmony Search (HS) algorithms were proposed to solve lot-streaming flow shop scheduling problem with setup time under both the wait and no-wait cases. The objective was to minimize the maximum completion time or makespan. Unlike the original HS algorithm, the proposed algorithm represented a harmony as a discrete job permutation and applied inver-over and reconstruction operators to generate a new mechanism. An efficient initialization scheme based on the modified Nawaz-Enscore-Ham (NEH) heuristic was presented to construct an initial population with a certain level of quality and diversity. A simple but effective local search was added to enhance the intensification capability. The computational results and comparisons show that the proposed algorithm is effective and efficient in finding better solutions for the problem considered.

Concerning the girth problem of LowDensity ParityCheck(LDPC) codes, a new rapid algorithm for detecting the girth of LDPC codes in combination with Dijkstra algorithm and the feature of Tanner graph was proposed, and the time complexity of this algorithms was lower. Compared with known algorithms, this algorithm not only can calculate rapidly, but also return the girth and edges only one time, thus avoiding redundant computation. At last, the simulation verifies the feasibility and efficiency of this new algorithm.