The Proposal of Pipeline for Photorealistic 3D Object Modeling

 

 

Qi Zhang

Haitao Jiang, Yun Ji, Linwei Fu, Tae Soo Yun

 

Department of Visual Contents

Graduate School, Dongseo University

 Busan, South Korea

E-mail: zoveht@gmail.com, jiang5275@gmail.com

jiyunmyself@naver.com, flw0819@gmail.com, yntaesoo@gmail.com

 

 

Abstract

 

In this article, an extremely simple method is used to make digital models. In this production process, firstly the object which needs to be produced is shot by a camera. Then a variety of 3D tools will be used to process, and finally a group of models and textures can be obtained. This article also shows the digital models produced by the author using this method, together with Agisoft, Warp 3, Maya and other production tools. And based on the contents hereby, the advantages and disadvantages of this production method are summarized.

 

Keywords-3D Modeling; Topology; Maya

 

1. Introduction

 

Nowadays, digital technology has become been much more developed than before, and the digital models production has reached to a very high level. However, there is still much room for improvement in production efficiency. There are numbers of 3D production tools used to improve production efficiency. This paper also adopts a new production method, which is more convenient than the production method used before. Meanwhile this new method guarantees the quality of the model, and lays a solid foundation for later production.

 

2. Preparation for Shooting

 

The production method used this time, as shown in Fig. 1(a), is that the shooting is carried out in a room illuminated by a led fluorescent lamp and the brightness can be unified, by which shadow interference can be greatly reduced. Thus it can be ensured that the model texture will not be interfered by the shadow. As shown in Fig. 1(b), then a Canon 5D digital SLR camera with manual focus and a fixed-focus lens set to 55mm will be used to shoot the object which needs to be made into a digital model. As shown in Fig. 1(c), in order that the texture marking of each detail of the object can be accurately got, the camera should shot the object by 360-degree surrounding. Finally it can be ensured that every detail of the object can be seen clearly by the materials shot. When preparing shooting, the better the equipment is, the better the results will be [4][5].

 

(a) Shooting environment             (b) Camera Settings            (c) Shooting Method

Fig. 1 Preparation for shooting

3. Production Method and Production Procedures

In this study, the three essential tools are Agisoft, for processing photo materials, Maya, for making models, and Warp 3, for topologies, which can be reasonable used to quickly get a group of models and textures.

Fig.2 shows the whole workflow, which includes three parts: The first part is about how to deal with the materials obtained from shooting. The screened photos are imported into Agisoft tool to get a group of models and textures. But the model surface number is too large, and the UV is not regular; the second part is about the acquisition of the digital model. Refer to the photo materials and do modeling by manual. Although it takes a long time, a high-quality model of regular wiring can be obtained; the third part is the acquisition of the texture. Use the textures obtained before the topology by Warp 3 tool, and get a regular texture.

 

 

Fig.2 Pipeline of production method

 

4. Research Procedures and Results

 

The above method was previously applied to the production of digital character models. So in this study, it is wanted that the production method can be applied to the production of hard surface objects. As shown in Table 1, this study basically follows this workflow. The production content of each step, and the digital model pattern produced after this step is completed, are also recorded in this table.

 

4.1. Research Procedures

 

According to the method as described above, prepare a room that brightness can be unified and a camera for shooting. Due to lack of equipment, etc., it is impossible to shoot a 360-degree subject, so the materials are obtained by surrounding shooting from three angles of up, middle and down.

Screen the photo materials obtained shooting and import into the Agisoft tool for processing. Then a group of high-mode textures can be obtained. However, they cannot be used because the number of models obtained is too large and the wiring is irregular. So it is needed to use the Maya tool referring to photo materials and the obtained high poly model to do manual modeling. Thus a high quality model with a well-organized wiring can be obtained.

A group of models through the Agisoft tool is obtained, at the same time, a group of textures are also obtained. The texture quality is very high, but because the UV of the model is irregular, the texture is also irregular. It is therefore necessary to import the obtained model and the model produced in the Warp tool into the Warp 3 tool for topology. While in the topology, attention should be paid to align the two models. The better the models are aligned, the higher the quality of the texture will be. By this method, a group of regular textures can be got quickly.

 

Table 1. Pipeline Offer for Photorealistic 3D Prop Modeling

 

 

4.2. Combination and Results

 

In this study, combination is not involved. But if the shape of the object that needs to be made into a three-dimensional model is too complicated, the model needs to be divided into several parts to be topologically mapped and then merged. Otherwise, the new texture will lose a lot of details and the topology time will take longer, then the work efficiency will be reduced. It can be seen from the table, by this method, that a group of models and textures with a high degree of precision can quickly be obtained.

 

5. Conclusion

 

In this article, an extremely simple production method is used to obtain high-quality digital models and textures. This method requires the flexible use of three production software: Agisoft, Maya, and Warp 3. Agisoft is able to quickly acquire a group of models and textures, which although cannot be directly used, they can still play a crucial role in the future work. The obtained model can be used as a reference to make the newly-made regular wiring, which is much closer to the real object. And the obtained texture, by the topology tool, can be processed to a regular texture.

The shortage of this method is that there is no way to topology the model as it does a digital character model, so manual modeling is required and a lot of time will be taken. However, compared to other methods, the addition of a group of models as reference can save some time. And this production method, compared to other production methods, can save a lot of time in texture production, and promote work efficiency.

 

References

 

[1]    Paul Graham, Graham Fyffe, Borom Tunwattanapong, Abhijeet Ghosh and Paul Debevec, "Near-Instant Capture of High-Resolution Facial Geometry and Reflectance", ICT Technical Report No. ICT-TR-01-2015, May, 2015.

[2] Pascal Bérard, Derek Bradley, Maurizio Nitti, Thabo Beeler, Markus Gross, ACM Transactions on Graphics (TOG) - Proceedings of ACM SIGGRAPH Asia 2014 TOG Homepage Volume 33 Issue 6, Article No. 223, November 2014.

[3] Paul Graham, Borom Tunwattanapong, Jay Busch, Xueming Yu, Andrew  Jones,  Paul  Debevec,  and  Abhijeet  Ghosh, "Measurement-Based Synthesis of Facial Microgeometry", Volume 32 (2013), Number 2, EUROGRAPHICS 2013.

[4] Jiang Hai Tao, Ji Yun, Tae Soo Yun, “The Digital Actor Hologram Performance Project: Achieving the Photorealistic Facial Skin and Rendering”, Dongseo University, 2017.

[5] Zhang Qi, Fu Linwei, Jiang Haitao, Ji Yun, Qu Lin, Yun Taesoo, “Photoscan method for Achieving the Photorealistic Facial Modeling and Rendering”, Dongseo University, 2018.