基于手绘简约线条图的CAD指令生成数据

通过采集用户手绘的简约线条图数据，利用深度生成模型进行训练，算法可以根据手绘线条生成相应的CAD指令代码。这一数据适用于工业设计与制造等领域，用户能够通过手绘线条快速生成复杂的CAD指令，进而得到三维模型，从而增强设计灵活性和效率，解决手绘创作转化为三维CAD模型过程中繁琐的操作问题。在移动触控设备上绘制线条图，通过算法解析并生成对应的CAD指令，从而加速三维建模过程，可用于工业设计与制造领域。具体过程如下：（1）使用移动触控设备和相关绘图软件，根据指定的CAD模型人工制作手绘简约线条图。（2）数据处理：应用图像去噪、二值化、膨胀和腐蚀等算法去除噪声并增强图像质量。然后，使用图像编码器提取线条特征。公式为：F_sketch = Encoder_sketch(Image_sketch)，其中，F_sketch表示手绘简约线条图特征，Encoder_sketch表示预训练的图像编码器，Image_sketch表示手绘线条简约图。（3）模型构建：使用手绘简约线条图特征作为自变量，CAD指令为因变量，设计并搭建深度学习模型。通过公式提取CAD指令：CAD_commands = Decoder_lines(F_lines)，其中，CAD_commands表示CAD指令，Decoder_lines为深度学习模型。然后将生成的CAD指令解析成CAD中间模型和CAD模型。最后使用平均ACC和平均IoU评估整体生成质量。

This dataset collects simple hand-drawn line sketch data from users, and is trained with deep generative models. The trained algorithm can generate corresponding CAD command codes based on input hand-drawn lines.

This dataset is applicable to fields including industrial design and manufacturing. Users can quickly generate complex CAD commands via hand-drawn lines, and further obtain 3D models, thereby enhancing design flexibility and efficiency, and solving the cumbersome operational issues in the process of converting hand-drawn creations into 3D CAD models. Drawing line sketches on mobile touch devices, the algorithm parses and generates corresponding CAD commands, thus accelerating the 3D modeling process, which can be used in industrial design and manufacturing scenarios.

The specific process is as follows:
1. Use mobile touch devices and related drawing software to manually create simple hand-drawn line sketches based on specified CAD models.
2. Data processing: Apply algorithms such as image denoising, binarization, dilation and erosion to remove noise and enhance image quality. Subsequently, an image encoder is used to extract line features. The formula is: $F_{sketch} = Encoder_{sketch}(Image_{sketch})$, where $F_{sketch}$ represents the features of the simple hand-drawn line sketch, $Encoder_{sketch}$ denotes the pre-trained image encoder, and $Image_{sketch}$ refers to the simple hand-drawn line sketch.
3. Model construction: Take the features of the simple hand-drawn line sketch as the independent variable and CAD commands as the dependent variable to design and build a deep learning model. The CAD commands are extracted via the formula: $CAD_{commands} = Decoder_{lines}(F_{lines})$, where $CAD_{commands}$ represents the CAD commands, and $Decoder_{lines}$ is the deep learning model. Then, the generated CAD commands are parsed into CAD intermediate models and final CAD models. Finally, average ACC and average IoU are used to evaluate the overall generation quality.

搜集汇总

数据集介绍

特点

该数据集通过手绘简约线条图生成CAD指令，适用于工业设计与制造领域，包含12031条数据，涵盖CAD模型、手绘线条图及其特征，并通过深度学习模型进行训练。