# 3D Visualization for Defect Inspection This notebook demonstrates how to integrate the **MVTec 3D Anomaly Detection Dataset** into FiftyOne. The dataset contains high-resolution 3D scans of objects, including **point cloud data** and **RGB images**, which are useful for anomaly detection tasks. ![visual_inspection](https://cdn.voxel51.com/getting_started_manufacturing/notebook7/visual_inspection.webp) ## Learning Objectives: - Convert TIFF to PCD format for visualization in FiftyOne. - Create a Grouped Dataset in FiftyOne - Leverage FiftyOne for visualization and analysis. ### Key Features: - **3D Representation**: The dataset provides XYZ point cloud representations stored as TIFF files. - **RGB and Mask Images**: Each sample includes an RGB image and a corresponding segmentation mask. - **Anomalous and Normal Samples**: The dataset includes both normal and defective objects for anomaly detection research. - **Grouped Datasets in FiftyOne**: FiftyOne supports the creation of grouped dataset which contain multiple modalities To make this dataset compatible with FiftyOne, we need to **convert TIFF files into PCD (Point Cloud Data) format** for visualization. ## Converting TIFF to PCD for Visualization MVTec 3D provides **XYZ coordinate data stored in TIFF format**, which must be converted to **PCD format** to be visualized in FiftyOne. The function below: 1. Loads the TIFF file as a **NumPy array**. 2. Reshapes it into **Nx3 (XYZ points) format**. 3. Saves it as a **PCD file** using Open3D. Additionally, another function includes **color segmentation** from masks to highlight anomalies. The next three cells are for your reference, so you can take a look at how to convert from TIFF to PCD. For the TIFF example, please select one sample from the [MVTec 3D-AD dataset](https://www.mvtec.com/company/research/datasets/mvtec-3d-ad/). ## Creating a Grouped Dataset in FiftyOne FiftyOne allows creating **grouped datasets**, where multiple data modalities (e.g., RGB images, segmentation masks, and point clouds) can be linked together under a common identifier. This enables: - **Synchronized visualization**: Easily switch between different representations of the same object. - **Multi-modal analysis**: Combine insights from images, masks, and 3D data. This notebook demonstrates how to create a grouped dataset where each sample includes: - An **RGB image** - A **segmentation mask** - A **3D point cloud (PCD)** We will use `potato` object from the [MVTec 3D Dataset](https://www.mvtec.com/company/research/datasets/mvtec-3d-ad/), here are the modified subset of the dataset. You can download the subset here [Potato MVTec 3D](https://huggingface.co/datasets/pjramg/potato_mvtec3d). Please download the dataset to disk and add samples in FiftyOne as I am showing you in the following cell.

Note: Loading this dataset using the Hugging Face hub won’t work because it is not saved in FiftyOne format. ### Create the Dataset instance in FiftyOne ### Read the MVTec Directory (potato object) and creating pairs of metadata before grouping dataset in FiftyOne ### Creating the grouped dataset in FiftyOne