TechING: Towards Real World Technical Image Understanding via VLMs
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917223586594816 |
|---|---|
| author | Nadeem, Tafazzul Shangari, Bhavik Rai, Manish Gupta, Gagan Raj Modi, Ashutosh |
| author_facet | Nadeem, Tafazzul Shangari, Bhavik Rai, Manish Gupta, Gagan Raj Modi, Ashutosh |
| contents | Professionals working in technical domain typically hand-draw (on whiteboard, paper, etc.) technical diagrams (e.g., flowcharts, block diagrams, etc.) during discussions; however, if they want to edit these later, it needs to be drawn from scratch. Modern day VLMs have made tremendous progress in image understanding but they struggle when it comes to understanding technical diagrams. One way to overcome this problem is to fine-tune on real world hand-drawn images, but it is not practically possible to generate large number of such images. In this paper, we introduce a large synthetically generated corpus (reflective of real world images) for training VLMs and subsequently evaluate VLMs on a smaller corpus of hand-drawn images (with the help of humans). We introduce several new self-supervision tasks for training and perform extensive experiments with various baseline models and fine-tune Llama 3.2 11B-instruct model on synthetic images on these tasks to obtain LLama-VL-TUG, which significantly improves the ROUGE-L performance of Llama 3.2 11B-instruct by 2.14x and achieves the best all-round performance across all baseline models. On real-world images, human evaluation reveals that we achieve minimum compilation errors across all baselines in 7 out of 8 diagram types and improve the average F1 score of Llama 3.2 11B-instruct by 6.97x. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_18238 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TechING: Towards Real World Technical Image Understanding via VLMs Nadeem, Tafazzul Shangari, Bhavik Rai, Manish Gupta, Gagan Raj Modi, Ashutosh Computation and Language Computer Vision and Pattern Recognition Machine Learning Professionals working in technical domain typically hand-draw (on whiteboard, paper, etc.) technical diagrams (e.g., flowcharts, block diagrams, etc.) during discussions; however, if they want to edit these later, it needs to be drawn from scratch. Modern day VLMs have made tremendous progress in image understanding but they struggle when it comes to understanding technical diagrams. One way to overcome this problem is to fine-tune on real world hand-drawn images, but it is not practically possible to generate large number of such images. In this paper, we introduce a large synthetically generated corpus (reflective of real world images) for training VLMs and subsequently evaluate VLMs on a smaller corpus of hand-drawn images (with the help of humans). We introduce several new self-supervision tasks for training and perform extensive experiments with various baseline models and fine-tune Llama 3.2 11B-instruct model on synthetic images on these tasks to obtain LLama-VL-TUG, which significantly improves the ROUGE-L performance of Llama 3.2 11B-instruct by 2.14x and achieves the best all-round performance across all baseline models. On real-world images, human evaluation reveals that we achieve minimum compilation errors across all baselines in 7 out of 8 diagram types and improve the average F1 score of Llama 3.2 11B-instruct by 6.97x. |
| title | TechING: Towards Real World Technical Image Understanding via VLMs |
| topic | Computation and Language Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2601.18238 |