README.md

---
license: mit
tags:
  - computer-vision
  - image-classification
  - pytorch
  - insurance
  - claims-triage
  - grad-cam
pipeline_tag: image-classification
---

# ClaimSight — Vehicle Damage Triage Classifier (resnet50)

Binary image classifier that flags whether a submitted vehicle photo
shows visible damage (`00-damage`) or an intact vehicle (`01-whole`),
built as a **claims-triage decision-support tool** — not an autonomous
adjuster. Every flagged prediction is intended to route to a human
reviewer.

Full project, training code, and API: https://github.com/<your-username>/claimsight

## Intended use

- First-pass triage of policyholder-submitted claim photos, to route
  obviously-intact vehicles away from a manual review queue.
- **Not** a final claims-adjudication system. **Not** a repair-cost
  estimator. Every prediction should be reviewed by a human before any
  claim decision is made.

## How to use

```python
import torch
from huggingface_hub import hf_hub_download

# clone github.com/<your-username>/claimsight for src/model.py, src/preprocessing.py
from src.model import build_model
from src.preprocessing import preprocess
from src.dataset import eval_transform

checkpoint_path = hf_hub_download(
    repo_id="<your-username>/claimsight-damage-detection",
    filename="best_resnet50.pt",
)
model = build_model("resnet50", pretrained=False)
model.load_state_dict(torch.load(checkpoint_path, map_location="cpu"))
model.eval()

import cv2
image_bgr = cv2.imread("claim_photo.jpg")
image_rgb = preprocess(image_bgr)                      # same function used in training
tensor = eval_transform(image=image_rgb)["image"].unsqueeze(0)
probs = torch.softmax(model(tensor), dim=1).squeeze()
print({"00-damage": probs[1].item(), "01-whole": probs[0].item()})
```

## Training data

Car Damage Detection (Kaggle, `anujms/car-damage-detection`) — 2,300
images, binary folder labels only (no masks/bounding boxes), split
1,840 train / 460 validation, balanced within each split.

## Architecture & training

Transfer learning (resnet50, ImageNet-pretrained) in two phases: frozen
backbone with a fresh head first, then fine-tuning of the last block at
a reduced learning rate. `ReduceLROnPlateau` + early stopping on
validation loss. Full details, augmentation policy and reproducibility
notes: see the project README.

## Evaluation (real, on the held-out validation split)

| Metric | Value |
|---|---|
| Validation accuracy | 0.9435 |
| ROC-AUC (damage class) | 0.9858 |
| Recall — damage class (@ threshold 0.5) | 0.9565 |
| Precision — damage class (@ threshold 0.5) | 0.9322 |
| Confusion matrix (TN/FP/FN/TP) | 214/16/10/220 |

Recall on the `damage` class is the priority metric: a false negative
(damaged vehicle classified as intact) can wrongly close a legitimate
claim, while a false positive only costs one extra human review. A
recall-priority operating point was chosen by sweeping the decision
threshold: at threshold **0.25**, damage recall is
**0.9870** at precision **0.9080**
(vs. 0.9565 recall / 0.9322 precision at the default 0.5).

| Architecture | Val. accuracy | ROC-AUC |
|---|---|---|
| **resnet50** (this checkpoint) | 0.9435 | 0.9858 |
| efficientnet_b0 | 0.8870 | 0.9620 |


## Explainability

Grad-CAM (`pytorch-grad-cam`) on the last convolutional layer is used
as a **weak-localization** signal — a coarse heatmap of the region that
most influenced the decision. It is **not** pixel-level segmentation:
there is no mask ground truth in this dataset, so there is no IoU/Dice.
See the project repo's `outputs/gradcam/` for overlays on both correct
and incorrect predictions, including a documented shortcut-learning
check.

## Limitations

- Binary output only — no severity or damaged-part classification.
- Grad-CAM is weak localization, not segmentation.
- Modest dataset size (2,300 images, one source) — real domain-shift
  risk against a real insurer's photo distribution (different brands,
  angles, lighting, phone cameras).
- Not validated as a repair-cost estimator — triage signal only.
- Requires human review in every deployment path.

## Reproducibility

Fixed seed (`torch.manual_seed(42)`), deterministic cuDNN settings,
pinned `requirements.txt`. Metadata for this exact run:

```json
{
  "arch": "resnet50",
  "seed": 42,
  "device": "cuda",
  "phase1_epochs_ran": 8,
  "phase2_epochs_ran": 15,
  "best_val_loss": 0.14994667431582576,
  "checkpoint": "models\\best_resnet50.pt",
  "torch_version": "2.6.0+cu124",
  "trained_at_utc": "2026-08-01T22:51:16Z"
}
```