Reliable documentation of the extent and areas of vehicle damage is essential for completing the repair processes and conducting insurance assessments. Bounding boxes, which enable the measurement of damaged areas, are drawing much popularity, but they can sometimes lack accuracy. This study proposes an innovative object detection model, YOLOv10, which has been trained and fine-tuned on car damage dataset. Our novelty lies in the techniques added to this model, including generating a convex hull and performing Delaunay triangulation as the original release of YOLOv10 does not carry out segmentation task. Instead of using the ordinary bounding box structures, the hull-based enclosure structure is utilized, and a highly detailed triangular mesh is generated that facilitates accurate measurement of the damaged area. We are trying to show how YOLOv10 combined with image processing (our novelty) stands up to the already existing model YOLOv8. This methodology is compared to the segmentation and object detection model, YOLOv8 which has also been fine-tuned on the same dataset. The results obtained in using Intersection over Union (IoU) scores, where both models were employed, demonstrated how effectively the novel approach works for segmentation of dents and scratches of the objects. We also measure the damage area completed after the segmentation to confirm the correctness and accuracy of our approach. We showcased our method’s usefulness on a dataset of 316 vehicle images for YOLOv10 object detection model and YOLOv8 segmentation model. Moreover, these enhancements also increase assessment accuracy and processing speeds. Our approach is in line with the development in the field, exploring the applicability of advanced computer algorithms in the assessment of damage to vehicles. Considering the advanced geometric models, this research addresses a gap. The incorporation of these techniques demonstrates the prospects of tremendous improvement in accuracy and efficiency in the recording of damages which could enhance repairs and the assessment of insurance claims. We have gotten a Box mAP-50 of 0.741 and Box mAP-50:95 of 0.541for YOLOv10 and Box mAP-50 score of 0.803, Box mAP-50:95 score of 0.586, Mask mAP-50 score of 0.783 and Mask mAP-50:95 score of 0.505 for YOLOv8. An average IoU of 0.52 is observed for YOLOv8 and 0.49 for YOLOv10.