According to the characteristics of tattoo images and the insufficient ability of the Convolutional Neural Network (CNN) to extract the image features in the full connection layer, a tattoo image detection algorithm based on three-channel CNN was proposed, and three aspects of improvement work were carried out. Firstly, the image preprocessing scheme was improved for the characteristics of tattoo images. Secondly, a CNN based on three-channel fully connected layer was designed to extracted and index the features. The spatial information extraction ability of different scales was enhanced effectively, and the efficient detection of tattoo images was realized. Finally, the generalization ability of the algorithm was verified by two data sets. The experimental results on the NIST data set show that the proposed preprocessing scheme has a 0.17 percentage points increase of total correct rate and a 0.29 percentage points increase of correct rate for tattoo images than Alex scheme. Under the proposed preprocessing scheme, the proposed algorithm has obvious advantages on the standard NIST tattoo image set. The correct rate of the proposed algorithm reaches 99.1%, which is higher than 96.3%, the optimal value published by NIST; and 98.8%, obtained by traditional CNN algorithm. There is also a performance improvement on the Flickr data set.

The traditional AVL (Adelson-Velskii and Landis) tree programming has been faced with the problem of too much code, complex process and high adjusting ratio. To solve these problems, a unified rebalancing method was developed and a generalized AVL (AVL-N) tree was defined. The unified rebalancing method automatically classifies the type of the unbalanced node in AVL tree and uses a new way to adjust the tree shape without using standard rotations. AVL-N tree with relaxed balance allows the height difference between the right sub-tree and left sub-tree doesn't exceed N(N ≥ 1). When insertions and deletions have been performed in AVL-N tree, the height difference between the right sub-tree and left sub-tree of some nodes may be higher than N. At that time the unified rebalancing would be applied to rearrange the unbalanced node's descendants. The simulation results indicate that the adjusting ratio of AVL-N tree reduced significantly with N increasing, it is less than 4% for N=5 and less than 0.1% for N=13. The adjusting ratio of AVL-N tree is far below other classic data structures, such as red-black tree, and allows for a greater degree of concurrency than the original proposal.