## Introduction
The bacterial flagellum is a nano-machine known for its specialized cell-wall-mounted engine. This engine harnesses chemical energy to turn a filament, thereby allowing the bacterium to navigate towards nutrient-rich environments.

Cryogenic electron tomography (cryo-ET) offers a view of these motors inside cells. Samples are prepared by rapid freezing, preserving cells in non-crystalline ice.  These frozen samples are imaged by transmission electron microscopy (TEM) at different angles relative to the imaging beam; this "tilt-series'' is then computationally reconstructed into a tomogram of the cell’s volume.

![](https://upload.wikimedia.org/wikipedia/commons/thumb/1/14/Cryogenic_electron_microscopy_workflow.svg/1920px-Cryogenic_electron_microscopy_workflow.svg.png)

[Here](https://www.cellstructureatlas.org/6-2-flagellar-motor.html) is a video that walks through slices of the tomogram highlighting the different complexes that make up the bacteria. 

Sub-tomogram averaging is a technique used to enhance the signal-to-noise ratio and improve the resolution of macromolecular structures within sub-cellular volumes. The images below present an example of a set of sub-tomograms being layered and averaged to produce a higher resolution picture of the flagellar motor. The averaging process enhances the structural details common to the aligned sub-tomograms while reducing noise. The refined structure can be interpreted and analyzed to gain insights into the biological processes occurring within the cell.

Sub-tomogram averaging necessitates numerous images, generally identified manually. Below is an example of sub-tomogram averaging. The top images are averaged to produce the larger, clearer image below.

![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4640061%2F3b0a259fdbbb49740c4d8ec6420ba9a5%2Fsub_tomo.jpg?generation=1710191595195251&alt=media)

In this contest, your task is to build a machine learning model capable of processing 2D images of bacteria to identify the location of the flagellar motor. The data provided are slices of tomograms from different bacteria with a corresponding label that denotes the width, height, and center coordinates of the bounding box for the flagellar motor (if present).