Enhancing Identity Document Classification in KYC Processes: An Evaluation of the Bag-of-Visual-Words Model and Segmentation Impact
Bibliographic record
Abstract
The growth of online services, such as financial services, travel agencies, and e-government, has emphasized the importance of an efficient Know Your Customer (KYC) process.Efficient identity verification and document classification are crucial for KYC, such as ensuring the alignment of submitted identity documents with requirements, categorizing them accurately, and verifying their completeness within the KYC process.This article proposes the utilization of the bag-of-visual words (BoVW) model, which combines SIFT, k-means, and SVM techniques, to achieve accurate identity document classification without relying on geometry transformations.We observed that while segmentation significantly enhances accuracy during testing by eliminating irrelevant parts, its impact on the training phase appears to result in a drop in the model's performance.This drop in performance might be associated with segmentation during the training phase, where the removal of irrelevant parts might have caused the algorithm to have difficulty in identifying which features to disregard within the samples.This also implies that introducing imperfections such as blurred and low brightness samples into training dataset could potentially enhance the classification model.To test the theory, we compiled a dataset consisting of 8,400 samples, divided into 20 classes.This single compiled dataset was then used to generate three different kinds of datasets: USGM (an unsegmented dataset), SGM (a segmented dataset), and SGM2 (a segmented dataset where the subject of interest is clearly visible in the samples, serving as the training dataset).Three different testing is used: same-variant, cross-variant, and k-fold cross-validation.Our model demonstrates an average accuracy up to 97.2%, which remains relatively consistent across different types of testing.
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 | 0.005 | 0.010 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.002 |
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".