Team project recommendation: Automatic cartoon face generation-by Feng Nie

The internship at the Asian Research Institute has been nearly two months. At the end of August, we started the ASE course to cultivate our engineering knowledge and capabilities, the core of software engineering is to use the learning thinking to continuously consolidate practical capabilities. Therefore, the Team Project is an excellent opportunity for improvement. As a team project, I think it is necessary to take into account the feasibility, integrity, and maximize the satisfaction of user needs. Undoubtedly, the recent boom in face and meitu shows that image software has become increasingly popular. Based on this, I recommend "automatic cartoon face generation" as our topic.

Nowadays, users are pursuing apps that are interesting and personal, and tend to prefer image processing and video software. With the emergence of social networks, more and more portraits need to be set by users. To reflect their personal style, users also rely on processing software to make themselves more personalized. At the same time, they are pursuing a certain new trend. As a result, the appearance of face cute makes it frequently used by stars to ordinary users. However, face cute depends on the user's personal operations. For example, each part needs to be spelled by the user, and the final effect is merged. This production process is undoubtedly so "intelligent" today, it seems monotonous. Based on the lack of intelligence, we can make the process of generating an avatar as "Automated" as much as possible ".

Project Analysis: Image Processing-based software has become popular. If you can combine interesting and concise and fast image generation algorithms, the demand for this type of software is very wide, because users always prefer convenient operations. In addition, the tide of face cute such software is certainly more popular with users.

1 requirement:

Users can directly obtain the final cartoon face or sketch through our software by uploading their own selfie photos without the need to splice complex face parts.

2. Implementation Process:

First, the user can use PCA technology to identify the face part.

Then, use the existing face sketch algorithm (sketch) to generate the corresponding face effect based on the user's photos.

Secondly, we need to recognize all parts of the face and classify them by type, such as eyes (divided into big eyes, small eyes, laughter, and crying) to train classifiers, train the best matching results of cartoon pictures in each part

Finally, the cartoon image of each part of the classifier is put together to become the output of the user.

3 advantages:

While maintaining interest, it is very convenient for users to operate, and the final result is closer to the user himself, improving the convenience of such software.

4. Competitive Advantages:

Currently, this type of hot software, such as bitstrips, only stays on the user's own to find various face types. Therefore, it is similar to games such as Jigsaw puzzles and make-up many years ago. It only promotes the concept of Fine Arts, as a result, it is not integrated with the latest image and face technologies, making it impossible for users to be truly convenient. Therefore, we can fully promote this concept of combining experimental technology with art, so that this product is not only fashionable but intelligent, therefore, after successful cooperation with similar big and non-intelligent software, this kind of software can expand the database and become very market-oriented.

 

Through the above analysis, the automatic face generation software is complete and can be well evaluated for Windows Phone applications. using existing technologies for production can achieve good results, therefore, it can be used as a choice for the major assignments of this ASE course. Of course, because the time is still relatively long, our Group will also discuss and finally select the most suitable project for team development.

Team project recommendation: Automatic cartoon face generation-by Feng Nie