MDCA Plate: A Two Stage License Plate Recognition System with Channel Attention and Multi-Dilation Features
Bibliographic record
Abstract
Traffic imagery from the wild contains small plates, blur, glare, dirt, and occlusion, which makes license plate recognition difficult. We present MDCA Plate, a two stage system that detects plates with YOLOv8 and recognizes cropped plates with a PaddleOCR based recognizer enhanced by two modules. The first module is a channel attention block with batch normalization and Swish activation that strengthens channel reweighting and stabilizes optimization. The second module is a multi dilation extractor with three parallel convolutions that aggregate fine strokes and broad context followed by attention guided fusion and a one by one convolution. The pipeline is fully reproducible on CCPD2019 with automatic split generation, file name driven label conversion, deterministic cropping, OCR style label files, and scripted training. The detector converges quickly and reaches mAP at IoU 0.50 of about 0.995, so recognition is the main lever for improvement. Under a shared protocol with identical detector and crops, the combined recognizer with multi dilation and channel attention attains the highest end to end accuracy on the challenge split with 54.37, surpassing the baseline with 53.98 and the single module variants with 51.47 for channel attention and 53.09 for multi dilation, while remaining near ceiling on the base split with 99.87. Training directly on the challenge domain yields 73.87 on challenge, highlighting domain shift rather than capacity as the principal bottleneck.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.020 | 0.014 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".