COCO Dataset Preparation & Object Detection App Development

I am looking for someone experienced in computer vision and object detection to help organize and prepare a subset of the COCO dataset for an object detection project. The main goal is to create a structured dataset focusing on four challenging scenarios: Small Objects – images containing objects that occupy a small portion of the image and are difficult for detection models to identify accurately. Low Illumination – images where objects are affected by poor lighting, darkness, or low visibility. Partially Visible Objects – images where objects are partially occluded or only a portion of the object is visible. Similar/Class Confusion – images containing visually similar objects or classes that can easily be confused by an object detection model. Main Tasks Work with a subset of the COCO dataset. Analyze images and annotations to identify relevant examples for each category. Organize the dataset into the four categories above. Preserve the original COCO annotations and bounding boxes. Ensure that the selected images are correctly labeled and organized. Avoid unnecessary duplication between categories where possible. Provide a clean and structured dataset ready for training and evaluation with YOLO-based object detection models. The final dataset will be used to analyze the limitations of lightweight object detection models in challenging real-world scenarios. Please include examples of previous computer vision, object detection, or dataset preparation projects in your proposal.