Home    How it Works    Subscribe    Contact
| submit to reddit | Delicious |

Zoompy provides a totally new manner to explore objects seing its hidden details. Although it seems simple, there is a series of steps to create such experiments, with a lot of tech under the hood.

1. Photography

In the first step, photos with magnification are taken using a digital microscope and photos in the actual scale are taken using a conventional camera. Usually, experiments have just one photo in actual scale. However, in order to cover just an inch of the object with magnification it may be necessary dozens of photos. The entire experiment may has hundreds of microscope images.

In order to cover a greater number of details of a single scene, multiple photos of the same scene might be taken with different focus. These photos are combined as shown below.

2. Photo combination

In the second step, it is performed the boundary matching in order to generate a single scene from series of photos. In this step, the correct transformation (translation, rotation, scale) between each photo is found to combine them and get a single scene.

3. Brightness/color equalization

Even in a controlled environment there are brightness and color variance between multiple photos taken from the same scene. Therefore, the equalization process is essential to combine multiple photos in a proper manner. Below is shown the result of this process.

4. Actual scale matching

Finally, the scene composed with a series of photos taken using a digital microscope is matched with the photo of the object in actual scale.

5. Tile Generation

Now we have the all photographs combined in a single photo. However, this single photo may has hundreds or thousands of mega pixels. Its size may exceed hundreds of megabytes. In order to send to the user just the set of images in the region he or she is seeing, that single photo is segmented in series of tiles. Tiles are generated in different levels of detail. As the user zoom some region, the tiles are updated and replaced by others with greater detail. This approach is widely used by maps applications. Below are shown some tiles of the experiment "printed textbook'" of the same region in different level of detail.

6. Visualization

The server stores all tiles of each experiment. At the first moment, the visualization tool just show the photo of the object in actual scale and a minimal number of tiles in low detail. As the user zoom or navigate through the experiment, new tiles are requested and displayed.