In the industrial e-commerce landscape, creative designs such as banners and
posters are ubiquitous. Extracting structured semantic information from
creative e-commerce design materials (manuscripts crafted by designers) to
obtain design semantics represents a core challenge in the realm of intelligent
design. In this paper, we propose a comprehensive automated framework for
intelligently parsing creative materials. This framework comprises material
recognition, preprocess, smartname, and label layers. The material recognition
layer consolidates various detection and recognition interfaces, covering
business aspects including detection of auxiliary areas within creative
materials and layer-level detection, alongside label identification.
Algorithmically, it encompasses a variety of coarse-to-fine methods such as
Cascade RCNN, GFL, and other models. The preprocess layer involves filtering
creative layers and grading creative materials. The smartname layer achieves
intelligent naming for creative materials, while the label layer covers
multi-level tagging for creative materials, enabling tagging at different
hierarchical levels. Intelligent parsing constitutes a complete parsing
framework that significantly aids downstream processes such as intelligent
creation, creative optimization, and material library construction. Within the
practical business applications at Suning, it markedly enhances the exposure,
circulation, and click-through rates of creative materials, expediting the
closed-loop production of creative materials and yielding substantial benefits.
In the realm of industrial e-commerce, creative designs play a crucial role in attracting customers and enhancing brand exposure. However, extracting meaningful information from these design materials has been a challenging task for intelligent design processes. This article introduces a comprehensive automated framework that tackles this challenge, bringing together various components to intelligently parse creative materials.
The Material Recognition Layer
The material recognition layer serves as the foundation of the parsing framework. It incorporates different detection and recognition interfaces to identify key elements within creative materials. This includes not only the detection of auxiliary areas but also the identification of various layers present in the design. Moreover, it encompasses different algorithms like Cascade RCNN and GFL for precise detection and recognition.
The Preprocess Layer
After recognizing the different components of creative materials, the preprocess layer filters and grades them. This process aims to refine the creative layers and ensure that only high-quality materials are considered for further analysis. By eliminating irrelevant or low-quality content, this layer helps in improving the overall efficiency and effectiveness of the parsing framework.
The Smartname Layer
The smartname layer focuses on assigning intelligent names to the parsed creative materials. By leveraging advanced algorithms and techniques, this layer ensures that the names accurately reflect the content and purpose of each material. This intelligent naming enables easier categorization and organization of design assets, leading to improved searchability and accessibility.
The Label Layer
In order to provide a comprehensive understanding of creative materials, the label layer covers multi-level tagging. This means that materials can be tagged at different hierarchical levels, allowing for enhanced categorization and classification. By assigning relevant labels to each material, this layer enables efficient retrieval of specific design assets when needed.
Overall, the intelligent parsing framework presented in this article offers significant benefits to various downstream processes. It facilitates intelligent creation by providing structured semantic information, allowing designers to make informed decisions. Moreover, it aids in creative optimization by identifying areas of improvement and optimizing the design elements accordingly. Additionally, the framework contributes to the construction of a valuable material library, accelerating the closed-loop production of creative materials.
It is worth noting that the concepts discussed here exhibit a multi-disciplinary nature. The framework combines elements from computer vision, machine learning, and design principles to enable intelligent parsing of creative materials. By bridging these disciplines, the framework showcases the potential of interdisciplinary collaboration in solving complex challenges within the industrial e-commerce landscape. With further advancements and refinements, this framework holds promise for revolutionizing the design processes and optimizing creative outputs in the future.