DentiAsk: A VQA Benchmark for Multimodal Reasoning in Panoramic Dental Radiographs

Debesh Jha; Tapas Kumar Dutta; Roshan Paudel; Onkar Kishor Susladkar; Aashish Ghimire; Anupam Dhakal; Sabin Adhikari; Nand Kumar Yadav; Deepak Ranjan Nayak; Pravin Pawar; William C. W. Chen; Glenda Reynolds

arXiv · 2607.08790

DentiAsk: A VQA Benchmark for Multimodal Reasoning in Panoramic Dental Radiographs

Debesh Jha·Tapas Kumar Dutta·Roshan Paudel·Onkar Kishor Susladkar·Aashish Ghimire·Anupam Dhakal·Sabin Adhikari·Nand Kumar Yadav·Deepak Ranjan Nayak·Pravin Pawar·William C. W. Chen·Glenda Reynolds

View original source ↗Read source document ↗Export metadata ↓

Abstract

Accurate interpretation of panoramic dental radiographs requires the integration of multiple reasoning capabilities: detection, spatial localization, and quantitative assessment. Despite recent advances in multimodal learning, existing medical visual question answering (VQA) benchmarks do not fully capture this complexity, often reducing the task to simplified classification or templated queries. As a result, they provide limited coverage of the diverse reasoning processes required for clinically meaningful interpretation. We introduce DentiAsk, a large-scale dental VQA benchmark that pairs high-resolution panoramic dental radiographs with clinician-validated question-answer pairs spanning three reasoning tiers: descriptive recognition, spatial localization, and numerical quantification across three high-prevalence pathologies: periapical radiolucency (PARL), impacted teeth, and dental caries. DentiAsk comprises 1,000 high-resolution radiographs annotated with 10,000 expert-curated QA pairs. To our knowledge, it is the first dental VQA benchmark to unify categorical, spatial, and quantitative reasoning as separately scored tasks within a single evaluation framework. We benchmark 10 state-of-the-art vision-language models, including LLaVA-v1.5, LLaVA-v1.6, Qwen-VL, InternVL2, and LLaVA-Med, and find that models achieve stronger performance on descriptive queries, whereas they degrade sharply on spatial localization and counting, exposing limitations in compositional, multi-step reasoning. These findings reveal a gap between visual recognition and clinically meaningful reasoning, establishing DentiAsk as a challenging benchmark for advancing multimodal reasoning in medical imaging.

Explore related subjects

q-bio.QM

Keep this discovery

BibTeX RIS

Debesh Jha, Tapas Kumar Dutta, Roshan Paudel, Onkar Kishor Susladkar, Aashish Ghimire, Anupam Dhakal, Sabin Adhikari, Nand Kumar Yadav, Deepak Ranjan Nayak, Pravin Pawar, William C. W. Chen, Glenda Reynolds. 2026-06-29. DentiAsk: A VQA Benchmark for Multimodal Reasoning in Panoramic Dental Radiographs. https://arxiv.org/abs/2607.08790

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Biology-in-the-loop: Amortized Adaptive Hit Discovery in CRISPR Screens

Many biological discovery problems require experiments to be selected sequentially under constrained budgets. CRISPR screening is a prominent example, as exhaustive perturbation testing is often infeasible and candidate perturbations must instead be prioritized over multiple experimental rounds. Despite the importance of this problem, existing benchmarks for adaptive hit discovery remain limited in scale and diversity. Here, we introduce AssayBench-Loop, a large-scale benchmark for adaptive hit discovery comprising 1,389 CRISPR screens across five phenotype categories. Beyond enabling systematic evaluation, its scale makes it possible to learn acquisition strategies across historical experiments. Building on this resource, we introduce AssayLoop, a sequential experimental design framework combining AssayFormer, a transformer-based amortized acquisition policy trained across historical screens to adapt from experimental feedback, with LLM-derived biological priors through an adaptive handoff. In this view, completed experiments become training data for learning how accumulated evidence should guide what to test next, while LLMs provide prior biological knowledge to seed the search. We further introduce AssayLLM, showing that the same principle can be extended directly to an LLM through task-specific post-training. On temporally held-out screens, AssayLoop achieves a 5.67-fold enrichment over random selection and recovers 27.7% of hits after assaying approximately 5% of the candidate library, outperforming existing adaptive-design methods and standalone LLMs, and AssayFormer alone. Performance improves with increasing historical training data and transfers to phenotype categories excluded from training. These results demonstrate the value of learning acquisition policies across historical experiments and combining them with broad biological priors for efficient adaptive hit discovery.

q-bio.QM↗

arXiv2026

Multi-Task Bacterial Colony Detection and Classification Using YOLOv8 with Edge Optimization for Resource-Constrained Deployment

Manual counting and classification of bacterial colonies are critical yet labor-intensive tasks in microbiology, prone to human error particularly on densely populated plates. This work proposes a multi-task deep learning framework trained on the Annotated Germs for Automated Recognition (AGAR) dataset (18,000 images; 9,202 training / 3,067 testing) to automate Colony Forming Unit (CFU) enumeration and species classification. A custom multi-task CNN employing global regression served as the baseline, but demonstrated limited performance in clustered colony environments due to the absence of spatial localization. To address this, a YOLOv8 object detection architecture was adopted with high-resolution 1024x1024 inputs, enabling instance-level colony detection and label assignment. The model achieved a classification accuracy of 98.13% and a counting accuracy of 98.27% (within a 10-colony margin), demonstrating strong predictive capability. To bridge the gap between model performance and practical deployability, the trained model was optimized through unstructured and structured pruning, ONNX conversion, and reduced-precision inference (FP32, FP16, INT8). On a Raspberry Pi 4B, ONNX FP32 and FP16 variants offered the best balance between inference speed (~6.4s) and accuracy (MAE ~2.20). Unstructured pruning preserved predictive accuracy (MAE ~2.01) without runtime gains, while structured pruning resulted in significant accuracy degradation (MAE ~6.3), revealing the sensitivity of instance-level colony detection to architectural compression. These findings provide practical guidance for selecting optimization strategies in resource-constrained laboratory deployments.

q-bio.QM↗

arXiv2026

ADMET-EvO: a self-evolving scientific agent for sustained research across heterogeneous tasks

Scientific agents can move beyond automated model building by using accumulated evidence to revise both their questions and experimental strategies. The challenge is sustaining this adaptation across heterogeneous tasks without overfitting decisions to internal validation. Absorption, distribution, metabolism, excretion and toxicity (ADMET) prediction provides a demanding setting across diverse assays, datasets and chemical domains. We therefore developed ADMET-EvO, an evidence-gated agent that formalizes endpoints, generates falsifiable hypotheses and tests interventions across data, feature and model axes. It carries supported, rejected and inconclusive outcomes forward to guide each new cycle. Across the 22-task Therapeutics Data Commons (TDC) ADMET benchmark, ADMET-EvO achieved the highest task-normalized score of 96.77. Evidence-guided selection reduced cumulative fitting time by 72.2% within a predefined non-inferiority margin. It also formalized 43 toxicity-related tasks and constructed endpoint-specific predictors. Together, these results show how ADMET-EvO can accumulate evidence, revise its strategy and expand its research scope over time.

q-bio.QM↗