Fine‐grained semantic ethnic costume high‐resolution image colorization with conditional GAN

Abstract

Grayscale image colorization, especially for ethnic costume images, is highly challenging due to its rich and complex color features. The existing image colorization methods usually take the costume image as a whole in practical applications that lead to the ignorance of the semantic information of different parts of the costume. It is known that each part's color distribution of the ethnic costume is different. So, the color mapping of other parts is also diverse, which is determined by distinctive ethnic characteristics. This study introduces fine‐grained level semantic information and proposes a high‐resolution image colorization model for ethnic costumes targeting enhancement, inspired by semantic‐level colorization. The semantic information of different regions of ethnic costumes has a significant impact on the performance of the coloring task. Using Pix2PixHD as the backbone network, we create a new network architecture that maintains color distribution correspondence and spatial consistency of costume images using fine‐grained semantic information. In our network, we take the splice result of fine‐grained semantic for ethnic costume and grayscale image as the conditions and then feed them into the generative adversarial networks. We also discuss and analyze the influences of the grayscale channel and fine‐grained semantics on discriminator. Extensive experiments demonstrate that our method performs well compared with other state‐of‐the‐art automatic colorization methods.

Mining high utility pattern with negative items in dynamic databases

“…Problem definition: Incremental HUIs mining (iHUIM) 13 The problem is defined as: Suppose the database is…”

“…In contrast, there is no algorithm in the incremental database that can mine HUIs containing negative items. To this end, in this paper, the incremental algorithm EIHI 13 is further improved into an efficient pattern mining algorithm that can handle both positive and negative items without changing the data structure and key techniques, and the improved algorithm is named EIHI-N. To verify the effectiveness of the incremental list indexing structure proposed in this paper, further improvements are also made in EIHI-N by replacing the ordinary utility list in EIHI with II-list and naming the algorithm as EIHI-II-N. Similarly, the IncCHUI 41 algorithm can mine closed sets of HUIs in incremental databases and is consistent with EIHI in the way it handles incremental data, both requiring merging the utility list of the original database with the utility list of the added database after each mining.…”

“…Problem definition: Incremental HUIs mining (iHUIM) 13 The problem is defined as: Suppose the database is D , if , where N is a nonempty transaction sets, and N D , then D ' is called the update of database D . Let minutil, N , and H be the minimum utility threshold specified by the user, incremental data, and a set of HUI found in D .…”

“…Among the existing HUPM algorithms, the two‐phase based algorithm requires multiple dataset scans to find HUIs and generates too many candidate sets, which leads to poor scalability, Just like FUP‐HUI 13 ; Tree‐based HUPM algorithms such as HUSP‐NIV 14 use a more compact structure, but the algorithm takes more time and memory to process the tree structure; The array‐based HUPM algorithm can handle large amounts of data, but it does not perform well in cases of insertion and deletion, and is not suitable for dynamic storage, the EHNL 15 algorithm is one of them; the structure of the list‐based HUPM algorithm is easy to implement, but the expensive operation of list merging still causes the algorithm to become less efficient.…”

Automatic Gray Image Coloring Method Based on Convolutional Network

“…With the emergence of digital media technology and the popularity of the Internet, the animation industry has been greatly developed and advanced [ 1 – 3 ]. Animation works usually have two forms of expression, two-dimensional animation and three-dimensional animation, among which two-dimensional animation works have strong representation, character drawing and coloring are more natural, and three-dimensional animation works are not limited by the physical engine.…”